Earnings Quality and Earnings Management: An Empirical ...
Transcript of Earnings Quality and Earnings Management: An Empirical ...
Earnings Quality and Earnings Management: An Empirical
Analysis Of The Provision For Credit Loss On Trade
Receivables Amongst FTSE 350 Companies
Jonathon Butler
This dissertation is submitted in partial fulfilment of the requirements for the degree of
Master of Business in Accounting, Waterford Institute of Technology.
Research Supervisor: Mr. John Casey, FCA, MSc (Finance)
August 2012
__________________________________________________________________________
i
ABSTRACT
Amidst persistent global economic uncertainty, an on-going sovereign debt crisis
across Europe and an environment of elevated and deteriorating credit risk, companies
continue to face numerous challenges. One such challenge comprises the risk of credit
loss on trade receivables, which companies provide for through a specific provision.
Prior empirical research has documented extensively the existence of earnings
management through discretionary accruals. This cross-sectional study examines the
existence and determinants of abnormal provision for credit loss on trade receivables
in the context of both earnings quality and earnings management.
While traditional determinants of earnings management including capital market,
contractual, performance, governance and auditor related variables are examined, the
capital market response to instances of extreme abnormal provision for credit loss on
trade receivables is also considered. Consistent with prior earnings management
research, correlation and regression analyses are utilised to determine the extent of
relationships between abnormal provision and these variables.
The mean level of abnormal underprovision of -9.9% provides strong evidence of
provisioning practice at sharp variance with the current credit risk environment, while
the results of regression analyses provide strong evidence supporting the debt
hypothesis of positive accounting theory.
Increasing levels of gearing and increasing gross margins are identified as
significantly explaining abnormal underprovision, while strong evidence supporting
the mitigating effects of robust corporate governance structures on abnormal
underprovision is also identified, where the board of directors is comprised of an
increasing proportion of INEDs.
Further analysis confirms that those companies with extreme abnormal
underprovision experience a significantly inferior post financial year end stock price
performance relative to those companies with extreme abnormal overprovision.
This result suggests that capital markets, in identifying lower quality earnings and
discounting stocks where accounting abnormalities have been identified, may mitigate
the effects of earnings management activity.
___________________________________________________________________________
ii
ACKNOWLEDGEMENTS
The completion of this dissertation signifies the end of my formal third level
education. However, its completion and my four years at W.I.T. would not have been
possible without assistance, guidance and inspiration from so many people.
I am truly grateful to Mr. John Casey for his extensive input, forbearance, practical
advice and guidance, without which, completion of this dissertation would not have
been possible.
I extend sincere thanks to PricewaterhouseCoopers (Waterford) for their financial
support at both undergraduate level and in the completion of the MBS in Accounting
programme.
To all of the lecturing and support staff on both the MBS in Accounting and
BA (Hons) in Accounting programmes - I am forever grateful for your support, advice
and assistance in helping me during my four years at WIT.
To all of my fellow classmates – I wish to thank you for a wonderful year.
The experience has left me with great memories and I have been privileged to work
with a group of such talented and determined people.
To all who assisted in reviewing this dissertation – your efforts and support are
sincerely appreciated.
Finally, yet most importantly, I extend heartfelt thanks to my family: to Mum,
Dad and Joanne. You have remained steadfast in supporting me throughout the years,
helping me in all of my achievements to date. I will forever be grateful for your
support, guidance and prayers.
___________________________________________________________________________
iii
ETHICAL DECLARATION
Specifically excluded from Turnitin© upload – included in hard copy version.
_____________________________ _____________________________
Jonathon Butler Date
___________________________________________________________________________
iv
TABLE OF CONTENTS
Abstract ........................................................................................................................... i
Acknowledgements ........................................................................................................ ii
Ethical Declaration ........................................................................................................ iii
Table of Contents .......................................................................................................... iv
List of Figures ............................................................................................................... ix
List of Tables ................................................................................................................ xi
List of Appendices ...................................................................................................... xiii
List of Abbreviations .................................................................................................. xiv
Chapter One: Introduction ......................................................................................... 1
1.1 Introduction .............................................................................................................. 2
1.2 Research Rationale and Context .............................................................................. 3
1.3 Research Questions and Research Objectives ......................................................... 4
1.4 Research Methodology and Limitations .................................................................. 5
1.5 Dissertation Structure: Diagrammatic Overview ..................................................... 6
Chapter Two: Literature Review ............................................................................... 7
2.1 Introduction .............................................................................................................. 8
2.2 Introduction to Earnings Quality ............................................................................. 9
2.2.1 Earnings Quality as a Concept ......................................................................... 9
2.2.2 Measures of Earnings Quality .......................................................................... 9
2.2.3 Earnings Quality and Earnings Management ................................................ 10
2.3 Introduction to Earnings Management................................................................... 11
2.3.1 Earnings Management as a Concept .............................................................. 11
2.3.2 Differentiation between Earnings Management and Manipulation ............... 11
2.3.3 Accounting Policy Flexibility and Earnings Management ............................ 11
2.3.4 Impact from Earnings Management Activity ................................................ 12
2.3.5 Accrual Accounting, Earnings Management and Accounting Earnings ....... 12
2.4 Earnings Management – Theoretical Perspectives ................................................ 12
2.4.1 Introduction to Theoretical Perspectives ....................................................... 12
2.4.2 Signalling Theory and Higher Quality Firms ................................................ 13
2.4.3 Signalling Theory and Lower Quality Firms ................................................ 13
___________________________________________________________________________
v
Chapter Two: Literature Review (Continued)
2.4.4 Positive Accounting Theory and Earnings Management ............................... 14
2.4.5 Agency Theory and Earnings Management .................................................. 15
2.4.6 Agent Specific Motives for Earnings Management ....................................... 15
2.5 Conclusion ............................................................................................................. 16
Chapter Three: Literature Review........................................................................... 17
3.1 Introduction ............................................................................................................ 18
3.2 Earnings Management and the Financial Condition of a Firm .............................. 19
3.3 The Determinants of Earnings Management ......................................................... 19
3.3.1 Contracting Motives and Earnings Management .......................................... 20
3.3.2 Executive Level Compensation and Earnings Management ........................ 20
3.3.3 Debt Contract Motives and Earnings Management ...................................... 21
3.3.4 Earnings Expectations and Earnings Management ....................................... 22
3.3.5 Regulatory Motives and Earnings Management ........................................... 22
3.3.6 Corporate Governance Structures and Earnings Management ..................... 23
3.3.7 Audit Committee and Earnings Management ............................................... 24
3.3.8 Auditor Type, Auditor Remuneration and Earnings Management ............... 24
3.3.9 Revenue Manipulation, Deferred Revenue and Trade Receivables ............. 25
3.4 The Response of Capital Markets .......................................................................... 26
3.4.1 Earnings Quality and Stock Price Performance ............................................ 26
3.4.2 Equity Offerings, Earnings Management and Stock Price Performance ...... 27
3.4.3 Magnitude of Capital Market Based Earnings Management ........................ 27
3.5 Conclusion ............................................................................................................. 27
Chapter Four: Literature Review ............................................................................ 28
4.1 Introduction ............................................................................................................ 29
4.2 The Provision for Doubtful Receivables: Model Development ........................... 30
4.2.1 The Provision for Doubtful Receivables: McNichols and Wilson (1988) .... 30
4.2.2 The Provision for Doubtful Receivables: Lev and Thiagarajan (1993) ......... 30
4.2.3 Manipulation of Trade Receivables: Ricci (2011) ....................................... 31
4.3 The Provision for Doubtful Receivables – IASB Accounting Guidance ............. 31
4.3.1 IAS 39: Section 58 – 59 ................................................................................ 31
___________________________________________________________________________
vi
Chapter Four: Literature Review (Continued)
4.3.2 IFRS 7: Section 7.16 and 7.37 ....................................................................... 31
4.4 Literature Review Conclusion .............................................................................. 32
Chapter Five: Research Methodology ..................................................................... 33
5.1 Introduction ............................................................................................................ 34
5.2 Research Rationale ................................................................................................. 35
5.3 Research Questions ................................................................................................ 35
5.4 Research Objectives ............................................................................................... 36
5.5 Research Hypotheses for Testing ........................................................................... 36
5.5.1 Objective One: Hypothesis ........................................................................... 36
5.5.2 Objective Two: Hypotheses .......................................................................... 37
5.5.3 Objective Three: Hypothesis ......................................................................... 41
5.6 Research Approach ................................................................................................ 41
5.6.1 Dependent Variable: Earnings Quality and Earnings Management ............. 42
5.7 Sample Selection Process ...................................................................................... 42
5.7.1 Sample Selection Context ............................................................................. 42
5.7.2 Sample Selection Refinement ........................................................................ 43
5.7.3 Sample Selection Refinement – Specific Elimination Procedures ................ 43
5.8 Data Sources .......................................................................................................... 44
5.9 Data Measures Utilised .......................................................................................... 46
5.9.1 Primary Dependent Variable ......................................................................... 46
5.9.2 Additional Variables and Measures ............................................................... 47
5.10 Data Validity and Reliability ............................................................................... 47
5.11 Testing Procedures ............................................................................................... 48
5.12 Limitations ........................................................................................................... 48
5.13 Conclusion ........................................................................................................... 49
Chapter Six: Research Findings ............................................................................... 50
6.1 Introduction ............................................................................................................ 51
6.2 Analysis of Identified Outliers ............................................................................... 52
6.3 Magnitude of Abnormal Provision for Credit Loss on Trade Receivables ........... 52
6.3.1 Descriptive Statistics ...................................................................................... 52
6.3.2 Univariate Analysis: Simple Linear Regression Analysis ............................ 53
___________________________________________________________________________
vii
Chapter Six: Research Findings (Continued)
6.4 Determinants of Abnormal Provision for Credit Loss on Trade Receivables ....... 54
6.4.1 Compliance with Underlying OLS Regression Analysis Assumptions ......... 54
6.4.2 Descriptive Statistics: Independent Explanatory Variables .......................... 54
6.4.3 Univariate Analysis: Simple Linear Regression Analysis ............................ 56
6.4.4 Multivariate Analysis: Multiple Regression Analysis .................................. 58
6.4.5 H2 to H15: Summary Findings ....................................................................... 64
6.5 Stock Price Performance of Extreme Abnormal Providers .................................. 65
6.5.1 Descriptive Statistics ...................................................................................... 65
6.5.2 Multiple Regression Analysis: Stock Price Performance Significance ........ 66
6.6 Conclusion ............................................................................................................. 67
Chapter Seven: Discussion ........................................................................................ 68
7.1 Introduction ............................................................................................................ 69
7.2 Magnitude of Abnormal Provision for Credit Loss on Trade Receivables ........... 70
7.2.1 Provisioning Activity at Variance with Credit Risk Environment ............... 70
7.2.2 Increased Credit Delinquency: Downside Risk of Elevated Write-Offs ...... 70
7.2.3 Regulatory Considerations ............................................................................ 71
7.3 Determinants of Abnormal Provision for Credit Loss on Trade Receivables ....... 71
7.3.1 Capital Market Variables – Limited Evidence .............................................. 71
7.3.2 Contractual Variables – Significant Evidence ............................................. 72
7.3.3 Performance Variables – Significant Evidence ........................................... 73
7.3.4 Governance Variables – Significant Evidence ............................................. 74
7.3.5 Auditor Variables – Mixed Evidence ........................................................... 75
7.3.6 Non Hypothesised Factors ............................................................................ 75
7.4 Stock Price Performance of Extreme Abnormal Providers .................................. 76
7.4.1 Evidence Supporting the Efficient Market Hypothesis ................................ 76
7.4.2 Capital Markets: A Potential Instrument for Effective Regulation .............. 76
7.4.3 Non Hypothesised Factors ............................................................................ 77
7.5 Conclusion ............................................................................................................. 77
___________________________________________________________________________
viii
Chapter Eight: Conclusion ....................................................................................... 78
8.1 Introduction ............................................................................................................ 79
8.2 Research Questions, Research Objectives and Findings ...................................... 80
8.3 Limitations of Research ........................................................................................ 81
8.4 Recommendations for Practitioners ...................................................................... 81
8.4.1 Auditors ......................................................................................................... 81
8.4.2 International Accounting Standards Board – Standard Setters ..................... 82
8.4.3 Capital Market Participants ........................................................................... 82
8.5 Recommendations for Future Research ................................................................ 82
8.6 Conclusion ............................................................................................................ 83
References ................................................................................................................... 84
Appendices .................................................................................................................. 95
The total word count, excluding both figures and tables, is 16,495
___________________________________________________________________________
ix
LIST OF FIGURES
Chapter One:
1.1 Dissertation Structure: Diagrammatic Overview ..................................................... 6
Chapter Two:
2.1 Income Smoothing Hypothesis .............................................................................. 10
2.2 Signalling Theory and Earnings Management ....................................................... 13
2.3 Positive Accounting Theory and Earnings Management ....................................... 14
2.4 Agency Theory and Earnings Management ........................................................... 15
Chapter Three:
3.1 Stock Price Performance: Posting Earnings Manipulation .................................... 26
Chapter Five:
5.1 Capital Market Determinants of Earnings Management ....................................... 37
5.2 Contractual Determinants of Earnings Management ............................................. 37
5.3 Performance Related Determinants of Earnings Management .............................. 38
5.4 Governance Specific Determinants of Earnings Management .............................. 39
5.5 Auditor Related Determinants of Earnings Management ...................................... 40
Chapter Six:
6.1 Simple Linear Regression Equation ...................................................................... 53
6.2 Simple Linear Regression Equation ...................................................................... 56
6.3 Multiple Regression One Equation ....................................................................... 59
6.4 Multiple Regression Two Equation ...................................................................... 60
6.5 Multiple Regression Three Equation .................................................................... 61
6.6 Multiple Regression Four Equation ...................................................................... 63
6.7 Mean Stock Price Performance Post Financial Year End ..................................... 66
Appendix C:
C.1 FTSE 350 Sector Specific Composition Summary ............................................. 109
___________________________________________________________________________
x
Appendix D:
D.1 Alternative Multiple Regression One Equation ................................................. 112
D.2 Stock Price Performance Multiple Regression One Equation ........................... 114
D.3 Stock Price Performance Regression Two Equation ......................................... 115
D.4 Stock Price Performance Multiple Regression Three Equation ......................... 116
Appendix G:
G.1 Frequency Distribution and Bell-Curve (Multiple Regression One) ................. 134
G.2 Scatter Plot of Regression Residuals (Multiple Regression One) ...................... 134
___________________________________________________________________________
xi
LIST OF TABLES
Chapter Three:
3.1 Executive Level Compensation and Earnings Management ................................. 21
3.2 Corporate Governance Structures and Earnings Management ............................. 23
3.3 Auditor Type, Auditor Remuneration and Earnings Management ....................... 24
Chapter Five:
5.1 Capital Market Determinants of Earnings Management ...................................... 37
5.2 Contractual Determinants of Earnings Management ............................................ 38
5.3 Performance Related Determinants of Earnings Management ............................. 39
5.4 Governance Specific Determinants of Earnings Management ............................. 40
5.5 Auditor Related Determinants of Earnings Management ..................................... 41
5.6 Capital Market Response to Earnings Management ............................................. 41
5.7 Sample Selection Refinement Summary ............................................................... 44
5.8 Financial Year End Dates Summary Analysis ...................................................... 44
5.9 Data Sources for Data Measures ........................................................................... 45
Chapter Six:
6.1 Six Identified Outliers ........................................................................................... 52
6.2 Descriptive Statistics for Magnitude of Abnormal Provision ............................... 53
6.3 Simple Linear Regression Significance Statistics ................................................. 53
6.4 Descriptive Statistics for Categorical Independent Variables .............................. 55
6.5 Descriptive Statistics for Continuous Independent Variables .............................. 55
6.6 Simple Regression Analysis Results: Full Sample (N=204) ................................ 57
6.7 Simple Regression Analysis Results: Underproviders Only (N=138) .................. 57
6.8 Multiple Regression One Significance Statistics (N=204) ................................... 59
6.9 Multiple Regression One: Fourteen Hypotheses (N=204) ................................... 59
6.10 Multiple Regression Two Significance Statistics (N=204) ................................ 61
6.11 Multiple Regression Two: Seven Hypotheses (N=204) ..................................... 61
6.12 Multiple Regression Three Significance Statistics (N=138) .............................. 62
6.13 Multiple Regression Three: Fourteen Hypotheses (N=138) ............................... 62
6.14 Multiple Regression Four Significance Statistics (N=138) ................................ 63
___________________________________________________________________________
xii
Chapter Six: (Continued)
6.15 Multiple Regression Four: Four Hypotheses (N=138) ....................................... 63
6.16 H2 to H15: Summary Findings ............................................................................. 64
6.17 Descriptive Statistics for Stock Price Performance: Underproviders ................. 65
6.18 Descriptive Statistics for Stock Price Performance: Overproviders ................... 65
Appendix D:
D.1 Impact of Outliers upon Preliminary Regression Analysis ................................ 111
D.2 Alternative Multiple Regression One Significance Statistics (N=204) ............. 112
D.3 Alternative Multiple Regression One: Fifteen Hypotheses (N=204) ................. 113
D.4 Stock Price Performance Regression One Sig. Statistics (N=25) ...................... 114
D.5 Regression One: Stock Price Performance Hypotheses (N=25) ........................ 115
D.6 Stock Price Performance Regression Two Sig. Statistics (N=25) ...................... 115
D.7 Regression Two: Stock Price Performance Hypothesis (N=25) ........................ 115
D.8 Stock Price Performance Regression Three Sig. Statistics (N=25) ................... 116
D.9 Regression Three: Stock Price Performance Hypotheses (N=25) ..................... 117
Appendix G:
G.1 Variance Inflation Factor Analysis Results ....................................................... 135
G.2 Test for Multicollinearity: Continuous Independent Variables ......................... 136
___________________________________________________________________________
xiii
LIST OF APPENDICES
Appendix A: Personal Reflection .................................................................. 95
Appendix B: IFRS 7 and IAS 39 ................................................................... 98
Appendix C: Details of Final Sample Population ....................................... 102
Appendix D: Multiple Regression Analysis Data ....................................... 110
Appendix E: Disclosure Notes: Extracts from Annual Reports .................. 118
Appendix F: Dependent Variable Dataset .................................................. 120
Appendix G: Methodology Continued: OLS Regression Assumptions ...... 128
___________________________________________________________________________
xiv
LIST OF ABBREVIATIONS
CRSP - Centre for Research in Security Prices
EPS - Earnings Per Share
EQ - Earnings Quality
FTSE 100 - FTSE 100 Index on the London Stock Exchange
FTSE 350 - FTSE 350 Index on the London Stock Exchange
IFRS - International Financial Reporting Standards
INEDs - Independent, Non-Executive Directors
OLS - Ordinary Least Squares Regression
U.S. - United States of America
U.S. GAAP - United States Generally Accepted Accounting Principles
U.S. GAO - United States Government Accountability Office
U.S. SEC - United States Securities and Exchange Commission
VAT - Value Added Tax
VIF - Variance Inflation Factor
Chapter 1
Introduction
Chapter 1 Introduction
___________________________________________________________________________
2
CHAPTER ONE
INTRODUCTION
1.1 Introduction
This chapter outlines the justification for this study, explores its underlying rationale
and establishes its context. Key themes and concepts that underpin this study are
discussed, while its relevance to the current business environment is also considered.
The research approach and contribution of this study are also explored, while the
chapter concludes with a diagrammatic overview of the structure of the dissertation.
Chapter 1 Introduction
___________________________________________________________________________
3
1.2 Research Rationale and Context
The importance of trade credit as a method of corporate financing is widely
documented, with empirical evidence suggesting that trade receivables account for
between 5 to 30 per cent of the total assets of European companies (Van Der Wijst
and Hol, 2002). Post the 2008 financial crisis, both private and corporate credit risk
remain elevated. With the intensification of the sovereign debt crisis across Europe
during 2011 and 2012, both macroeconomic uncertainty and credit risk have increased
further. EFMA (2012) states that 79 per cent of credit risk manager respondents
anticipate a renewed recession across Europe during 2012, while restrictions in trade
credit and sharp increases in credit delinquencies are also anticipated.
Companies provide for anticipated losses on trade receivables through a specific
provision for credit loss on trade receivables (IAS 39: S.58-59), commonly referred to
as the provision for doubtful receivables or provision for bad debts. In an environment
of such elevated credit risk, any reduction in this provision or failure to augment the
provision relative to an increase in total gross trade receivables is suspect. Indeed, any
reduction or failure to augment this provision generally serves to inflate overall
earnings. This study examines such abnormal provisioning activity in the context of
both earnings quality and earnings management.
Prior empirical studies, ranging from Healy (1985), Jones (1991) to Dechow et al
(2011) have examined in great detail the existence, frequency and magnitude of
earnings management activity, primarily in the context of discretionary accruals.
While a limited number of recent studies including Chen (2006) examine earnings
management in an international context, the majority of prior research has been
conducted in a U.S. or U.S. GAAP compliant financial reporting context. McNichols
and Wilson (1988) identify the need for further research with regard to earnings
management through singular accrual measures such as the provision for bad debts.
However, excepting the research of Lev and Thiagarajan (1993) and Ricci (2011), the
researcher is not aware of additional extensive research that has examined the
manipulation of the provision for credit loss on trade receivables from an earnings
management perspective, in a European or IFRS compliant financial reporting
context.
Chapter 1 Introduction
___________________________________________________________________________
4
This study complements existing earnings management research in adopting
methodology employed in prior studies, while also adding to existing earnings
management literature. In considering the response of capital markets to instances of
extreme abnormal provision for credit loss on trade receivables, this study also
examines the link between theoretical perspectives and the real world business
environment. Finally, the results provide a series of useful information for
practitioners including auditors, accounting standard setters and capital market
participants, with regard to the magnitude and determinants of abnormal provision for
credit loss on trade receivables.
1.3 Research Questions and Research Objectives
The research questions to be addressed in this study are:
RQ 1 – What is the magnitude and what are the determinants of abnormal provision
for credit loss on trade receivables amongst FTSE 350 companies?
RQ 2 – What is the capital (stock) market response to instances of extreme abnormal
provision for credit loss on trade receivables amongst FTSE 350 companies?
Utilising abnormal provision for credit loss on trade receivables as a measure of
earnings quality and as a proxy for earnings management activity, the following
research objectives support the investigation of the research questions:
1. To quantify the existence, direction and magnitude of abnormal provision for
credit loss on trade receivables amongst FTSE 350 companies.
2. To develop a multivariate OLS regression model that examines the
applicability of previously identified and alternative determinants of earnings
management activity, including capital market, contractual, performance,
governance and auditor related variables to abnormal provision for credit loss
on trade receivables amongst FTSE 350 companies.
3. To examine the individual and aggregate stock price performance of the most
extreme abnormal providers for credit loss on trade receivables (both under
and over providers) over a specified post financial year end period.
Chapter 1 Introduction
___________________________________________________________________________
5
1.4 Research Methodology and Limitations
Consistent with prior earnings management research, including Frankel et al (2002)
and Chen (2006), this study identifies a core measure of earnings quality and proxy
for earnings management, subsequently utilising descriptive statistics, correlation and
regression analyses to complete the varying tests that examine the research questions.
All data is gathered directly from both the latest available annual reports and the
Thomson One Banker database. The final sample population consists of 204 FTSE
350 companies, while the primary measure of earnings quality and proxy for earnings
management comprises the relative change in the provision for credit loss on trade
receivables after controlling for the relative change in total gross trade receivables.
Limitations of this study relate to the potential omission of earnings management
through alternative abnormal provisioning or discretionary accruals. The attachment
of a single earnings inflation motive to abnormal underprovision also disregards the
possibility that overprovision may represent earnings deflation activity. This study
also assumes that the prior year provision for credit loss on trade receivables is
representative of steady state, un-managed, normal provisioning. It is possible that
provisioning activity was elevated during 2008 and 2009, amidst the core of the
financial crisis, with companies now reducing their provision for credit loss on trade
receivables in subsequent years.
Chapter 1 Introduction
___________________________________________________________________________
6
1.5 Dissertation Structure: Diagrammatic Overview
This dissertation consists of eight chapters in total. The overall structure of the
dissertation and composition of each chapter is detailed in Figure 1.1 below.
Figure 1.1 – Dissertation Structure: Diagrammatic Overview
Introduction
Chapter 1:
An overview of the context and rationale
underlying the study. The research questions
and research objectives are also introduced.
Literature
Review
Chapter 2 – 4:
An extensive examination of theory and
prior empirical research underlying earnings
quality and earnings management activity.
Methodology
Chapter 5:
A thorough overview of the data collection
and analysis procedures, with a detailed
consideration of the limitations of the study.
Findings
Chapter 6:
Descriptive statistics and the results of
correlation and regression analyses
undertaken are presented sequentially.
Discussion
Chapter 7:
The implications of the findings are
analysed relative to regulation, theory and
prior empirical research.
Conclusion
Chapter 8:
The major findings of the dissertation are
presented, along with recommendations for
both practitioners and future research.
Chapter 2
Literature Review:
Earnings Quality and Earnings Management:
Introduction and Theoretical Perspectives
Chapter 2 Literature Review
___________________________________________________________________________
8
CHAPTER TWO
LITERATURE REVIEW
Earnings Quality and Earnings Management:
Introduction and Theoretical Perspectives
2.1 Introduction
The purpose of this chapter is to provide an introduction to the concepts of earnings
quality and earnings management. Initially discussing prior and current literature on
these concepts, the chapter then examines both accounting and economic theory
underlying earnings management practice, including positive accounting theory,
signalling theory and modern theory of the firm.
Chapter 2 Literature Review
___________________________________________________________________________
9
2.2 Introduction to Earnings Quality
2.2.1 Earnings Quality as a Concept
The concept of earnings quality is documented extensively in prior literature, with
Ayres (1994) stating that earnings quality was examined as early as the 1930’s,
whereby the true or underlying value of a security could be determined through
careful analysis of an entity’s financial statements to indicate whether a company
should be trading in excess of or below its current market value. According to Ayres
(1994), a focus on the degree of permanence in reported earnings became a principal
measure of earnings quality during the early 1970’s.
Bricker et al (1995) posit that reported earnings are of the highest quality when they
are most reflective of underlying events and conditions. Moreover, Duncan (2002)
asserts that management must often undertake subjective estimates with regard to
losses on loans or trade receivables that directly impact earnings quality and that if
managers smooth or manage earnings through estimates that are either too liberal or
conservative, there is a significant risk that such earnings may be viewed as lower
quality earnings by financial statement users. In supporting Duncan (2002), Schipper
et al (2003) determine that investors consistently attach higher price multiples to
earnings patterns that are supported by high quality earnings and that the magnitude
of any earnings management activity directly impacts the quality of earnings.
2.2.2 Measures of Earnings Quality
Schipper and Vincent (2003) state that earnings quality may be measured through
indicators that include the ratio of cash from operations to income, changes in total
accruals or the direct estimation of discretionary accruals through accounting
fundamentals. Palliam and Shalhoub (2003) define earnings quality as a measure of
the predictability of future earnings while Schipper et al (2003) also posit that higher
quality earnings have a signalling effect that indicates the sustainability of an earnings
pattern. Bellovary et al (2005), in identifying the provision for doubtful receivables as
a singular measure of earnings quality, also contend that earnings quality refers to the
stability, persistence and lack of variability in reported earnings while Mohammady
(2010) supports these assertions in stating that persistence, predictive value, feedback
value and earnings smoothness can all be employed as indicators of earnings quality.
Chapter 2 Literature Review
___________________________________________________________________________
10
Figure 2.1 below, adapted from Ayres (1994) outlines the traditional income
smoothing hypothesis, where long run reported and managed or manipulated earnings
are smooth, relative to underlying, real earnings.
Figure 2.1 – Income Smoothing Hypothesis
2.2.3 Earnings Quality and Earnings Management
As a result of the relationship between earnings management activity and earnings
quality, the detection of earnings management has been the focus of multiple
empirical studies to date, with the work of Healy (1985), DeAngelo (1986), and Jones
(1991) in the development of specific models to test for the existence, frequency and
magnitude of earnings management. Subsequent empirical studies, including those of
Sweeney (1994), Dechow et al (1995), Dechow and Dichev (2002), and Dechow et al
(2011) have attempted to refine prior models in the detection of earnings management
and to identify the primary determinants of earnings management practice generally.
Chapter 2 Literature Review
___________________________________________________________________________
11
2.3. Introduction to Earnings Management
2.3.1 Earnings Management as a Concept
References to earnings management are not always explicit and are often described as
earnings manipulation, big bath accounting, income smoothing or creative accounting
(Stlowy and Breton, 2004). Dechow and Skinner (2000) suggest that there is a
somewhat limited degree of empirical evidence from academic studies to suggest that
earnings management has a material impact upon average reported earnings. Dechow
and Skinner (2000) also acknowledge that there is significant disparity between
academic and practitioner perspectives, whereby academics focus primarily upon
earnings management activity driven by contractual agreements, yet practitioners are
primarily concerned with capital market determinants of earnings management.
2.3.2 Differentiation between Earnings Management and Manipulation
Earnings management comprises income smoothing behaviour but also refers to the
intentional structuring of disclosure or investment decisions with the bottom line
impact in mind (Ayres, 1994). In defining earnings management, an important
distinction must be made between earnings management and earnings manipulation
more specifically. Dechow and Skinner (2000) suggest that while underprovision for
bad debts constitutes aggressive earnings management, it is fictitious inventory and
revenue inflation that constitute fraudulent earnings manipulation. Moreover, Chen
(2006) posits that expense recognition deferral, erroneous revenue recognition and
measurement abuse constitute outright earnings manipulation.
2.3.3 Accounting Policy Flexibility and Earnings Management
While acknowledging that accounting flexibility is a primary mechanism through
which earnings management takes place, Dechow and Skinner (2000) suggest that the
elimination of all accounting flexibility would render earnings useless as a
measurement of economic performance. Colson et al (2010) suggest that firms may
utilise such flexibility to provide a clearer indication of their financial performance,
rather than to mislead investors. Additionally, Srivastava (2008) determines that firms
utilise flexibility in revenue recognition rules in order to convey value relevant
information to investors and not to engage in earnings management.
Chapter 2 Literature Review
___________________________________________________________________________
12
2.3.4 Impact from Earnings Management Activity
Despite an apparent lack of conclusive empirical evidence within academic literature
(Healy and Wahlen, 1998), the regular occurrence of corporate earnings scandals
including those of Enron, Tyco and Global Crossing (Bitner, 2005) provides
supporting evidence that earnings management occurs in extreme forms with
significant adverse impacts on firms and their respective stakeholder groups.
Additionally, Healy and Wahlen (1998) state that there is evidence of significantly
negative stock market responses to allegations of earnings management, with a
corresponding risk of an adverse impact on resource allocation in the wider economy.
2.3.5 Accrual Accounting, Earnings Management and Accounting Earnings
Accrual accounting, which is utilised to disrupt cashflow patterns in order to
compensate for issues of both timing and recognition (Dechow and Dichev, 2002),
is most contentious in the area of earnings management. Although accounting driven
accruals are often identified as a primary mechanism through which earnings
management may take place (Dechow et al, 1995), it is argued that the use of accrual
accounting in the determination of earnings results in long run earnings patterns that
are closely correlated with returns (Degeorge et al 1999). Healy and Wahlen (1998)
also contend that current earnings, which are indicative of management judgement,
are value relevant and are better indicators of future cash flow performance than
current cash flows.
2.4. Earnings Management – Theoretical Perspectives
2.4.1 Introduction to Theoretical Perspectives
The determinants of earnings management activity can be viewed primarily within the
confines of the signalling, agency and positive accounting theoretical frameworks.
While several sources of literature contend that there are significant differences
between agency theory and signalling theory, Morris (1987) concludes that both
theories are consistent, with considerable overlap in many instances.
Chapter 2 Literature Review
___________________________________________________________________________
13
Figure 2.2 – Signalling Theory and Earnings Management
2.4.2 Signalling Theory and Higher Quality Firms
Spence (1973) suggests that where two parties are engaged in a transaction and there
exists the problem of asymmetric information, one party may send a signal to the
other, in order to convey value relevant information, with resultant positive
implications for the party sending the signal from a valuation perspective. Accounting
standard guidance prescribes a lower bound or minimum information disclosure
requirement level according to Morris (1987), who posits that higher quality firms
will utilise accounting information disclosure to indicate to shareholders that they are
not utilising accounting flexibility to their detriment, or that they are not utilising such
flexibility to the same extent as other firms.
2.4.3 Signalling Theory and Lower Quality Firms
Conversely, Morris (1987) also argues that lower quality firms who are determined
that accounting standard disclosure requirements do not provide fine information
signals; will engage in corporate lobbying to ensure that standards of this kind are
introduced. Morris (1987) further posits that, in the context of accounting policy
choice, higher quality firms will chose more optimal accounting policies that reveal
their superior quality when compared with lower quality firms, who will utilise
accounting methods that conceal their inferior quality. The clear inference from these
assertions is that those firms who engage in earnings management are predisposed
towards those signalling motives of inferior quality firms.
Signalling
Theory
Optimal Accounting
Policies Revealing
Superior Quality
Higher Quality Firms
Lower Quality Firms Accounting Methods
Concealing Inferior
Quality
Chapter 2 Literature Review
___________________________________________________________________________
14
Figure 2.3 – Positive Accounting Theory and Earnings Management
2.4.4 Positive Accounting Theory and Earnings Management
Developed primarily by Watts and Zimmerman (1978), positive accounting theory
provides significant grounding for the determinants of earnings management practice.
Watts and Zimmerman (1986) posit that, in the absence of manipulation by
management, earnings otherwise follow a particular process and in order to reduce the
variance of that process, management adopt or alter specific accounting procedures.
In analysing the relationship between earnings and stock prices, Watts and
Zimmerman (1986) also suggest that the method of generating reported earnings has
an important bearing upon the income smoothing hypothesis, whereby, all else being
equal, managers will smooth earnings.
Watts and Zimmerman (1986) further determine, from prior empirical evidence, that
although there is no definitive evidence of a relationship between the capital intensity
of a firm, political costs and earnings deflation, there is consistent evidence that key
variables including size, the debt to equity ratio and the existence of a management
level compensation plan impact the propensity towards earnings management within a
firm. Watts and Zimmerman (1986) also posit that there is a positive relationship
between an increasing debt to equity ratio, the existence of a management level
compensation plan and the likelihood of the adoption of earnings inflation procedures
specifically. Consequently, these variables have been subject to widespread testing in
earnings management research.
Increased Propensity
Towards Earnings
Inflation Activity
Management
Compensation
Hypothesis
Debt Hypothesis Positive Relationship
Between Debt and
Earnings Inflation
Positive
Accounting
Theory
Chapter 2 Literature Review
___________________________________________________________________________
15
Figure 2.4 – Agency Theory and Earnings Management
2.4.5 Agency Theory and Earnings Management
Conceptualised primarily by Jensen and Meckling (1976), agency theory defines the
relationship between the principal and agent, owner and manager of a firm. Palliam
and Shalhoub (2003) state that the risk differential between principals and agents
creates a problem in the principal - agent relationship. While the responsibility for the
management of earnings rests with the agents of a firm, the methods undertaken to
manage earnings are not equally desirable from a principal’s perspective (Palliam and
Shalhoub, 2003). The principal can limit divergence by the agent from their desired
perspectives through both incentives and monitoring costs. However, where both
parties strive for utility maximisation, divergence remains highly likely (Jensen and
Meckling, 1976).
2.4.6 Agent Specific Motives for Earnings Management
In examining revenue recognition practice in the context of an agency setting, Dutta
and Zhang (2000) determine that no performance measure based upon current
accounting information will result in optimal agent specific incentives where mark to
market accounting is utilised. Moreover, in order to comply with consensus earnings
forecasts, the desires of the principal or to project a smooth earnings path, Palliam and
Shalhoub (2003) posit that agents will manage earnings through the acceleration or
deferral of either revenue or expenses or through accounting operations.
Conflicting From
Agent and Principal
Perspectives
Methods of Earnings
Management
Agent Motives Compliance with
Expectations of
Principal & Markets
Agency
Theory
Chapter 2 Literature Review
___________________________________________________________________________
16
2.5 Conclusion
Clearly, the concepts of earnings quality and earnings management have been the
subject of extensive focus over several decades from both empirical and theoretical
perspectives. There is considerable cross literature consensus with regard to earnings
quality measures ranging from persistence to smoothness, along with the use of
singular measures of earnings quality relating to trade receivables as outlined by
Bellovary et al (2005). Additionally, accounting and economic theory provides strong
support for the existence and determinants of earnings management activity, with
many of these determinants examined extensively in the following chapter.
Chapter 3
Literature Review:
The Determinants of Earnings Management:
Varying Perspectives and the Response of
Capital Markets
Chapter 3 Literature Review
___________________________________________________________________________
18
CHAPTER THREE
LITERATURE REVIEW
The Determinants of Earnings Management:
Varying Perspectives and the Response of Capital Markets
3.1 Introduction
The purpose of this chapter is to discuss extensively the determinants of earnings
management practice as identified in prior empirical studies, while simultaneously
aiming to demonstrate a clear link between the prior mentioned theory and extant
earnings management practice. Where multiple authors’ findings relating to the
determinants of earnings management are discussed; summary tables are presented.
The chapter then concludes with a focus on the response of capital markets to
earnings management activity.
Chapter 3 Literature Review
___________________________________________________________________________
19
3.2 Earnings Management and the Financial Condition of a Firm
While earnings management more frequently takes place within firms that are
experiencing financial distress (Chen, 2006), healthy firms that have not experienced
multi-period accumulated losses also engage in earnings management
(Peltier-Rivest and Swirsky, 2000). However, the determinants of earnings
management within healthy firms are not equivalent to those of distressed firms, as
such traditional determinants, including executive level motives, would not yield
sufficient benefits in order to influence managers’ accounting choice (Peltier-Rivest
and Swirsky, 2000).
Jeffrey et al (2008) determine that where a firm has suffered significant prior period
operating losses or negative cash flows, it will be motivated to manipulate revenues
specifically rather than earnings generally, as capital market participants tend to value
such firms on the basis of the level and growth in revenue rather than earnings and
cash flows. Management within healthy firms that are engaged in union or labour
negotiations are more likely to make income decreasing total accruals that depress
total earnings (Peltier-Rivest and Swirsky, 2000). Moreover, within healthy firms,
there is limited evidence of strong earnings management incentives driven through
top level executive change or government lobbying – that is, when a firm is subject to
governmental investigation (Peltier-Rivest and Swirsky, 2000).
3.3 The Determinants of Earnings Management
In establishing the determinants of earnings management, a primary consideration
includes the extent to which such activity is driven by the exercise of managerial
judgement (Healy and Wahlen, 1998). Bitner (2005) contends that asset quality, sales
growth and depreciation contain key indicators highlighting some of the contributing
factors to earnings management, while also contending that, inter alia, the following
components contribute to earnings management activity:
Sub optimal decision making resulting from prolonged periods of prosperity with
warning signals being disregarded and contradictory evidence being rationalised.
Fear on the part of lower level executives who act in their own self-interest in the
non-disclosure of negative information to their superiors.
A deficit in longer term planning, operational viability and leadership.
Chapter 3 Literature Review
___________________________________________________________________________
20
3.3.1 Contracting Motives and Earnings Management
Contracting motives for earnings management refer to those determinants within the
confines of the agency relationship, where covenants or provisions are utilised to
mitigate traditional agency problems (Jensen and Meckling, 1976). Healy and Wahlen
(1998) state that contracting motives arise where management compensation contracts
are utilised to align external stakeholder and management incentives or where lending
contracts are utilised to prevent against managerial level engagement in activity to the
detriment of a firm’s creditors.
3.3.2 Executive Level Compensation and Earnings Management
According to Healy and Wahlen (1998), the balance of empirical evidence suggests
that managers will utilise accounting judgement to inflate earnings where bonus plans
and contractual compensation incentives are indexed to earnings performance.
Watts and Zimmerman (1986) suggest that managers with contractual bonus plans are
more likely to adopt accounting policies that lead to the premature recognition of
future period earnings in the current accounting period. Moreover, Healy (1985)
determines that firms who specify a limit on their bonus award schemes are more
likely to report accruals resulting in the deferral of income when the bonus limit is
reached, indicating that there is an incentive to report earnings that will result in
receipt of the maximum bonus level, but not beyond such a level.
Dechow and Sloan (1991) determine that chief executive officers reduce research and
development spending during their final years in office, possibly to report more
positive short run earnings, with their final compensation contracts linked to these
earnings upon departure. However, Healy and Wahlen (1998) contend that such
changes in research and development expenditure may arise as a result of changes in
general investment policy rather than earnings management specifically.
Siagian (2002) finds no evidence of an abnormally high association between the
bonus of the chief executive officer and annual earnings amongst firms have been
subject to enforcement actions by the U.S. S.E.C. However, Balsam (1998)
determines that management engage in accounting choice in order to enhance their
level of compensation, determining that the association between the compensation of
the chief executive officer and the reported income of a firm increases with the level
of discretionary accruals.
Chapter 3 Literature Review
___________________________________________________________________________
21
Gaver and Gaver (1998) support the findings of Balsam (1998), determining that
managers are rewarded for undertaking accounting choice that positively impacts
income. Chen (2006) also asserts and finds, in a Taiwanese context, that those firms
engaged in earnings manipulation have a stronger intention to avoid reporting net
losses or depressed earnings in order to secure high levels of bonus payments, when
compared with a sample of non-manipulating firms. Clearly, the majority evidence
from these studies (Table 3.1 below) suggests that executive level compensation is a
primary determinant of earnings management activity.
Table 3.1 – Executive Level Compensation and Earnings Management
Author Subject Earnings Management R
Healy (1985) Bonus Plans Limited Earnings Mgmt. +
Healy and Wahlen (1998) Bonus Plans Earnings Inflation +
Dechow and Sloan (1991) R&D Expenditure Earnings Inflation -
Balsam (1998) Mgmt. Compensation Earnings Inflation +
Gaver and Gaver (1998) Mgmt. Compensation Earnings Inflation +
Chen (2006) Bonus Payments Earnings Inflation +
Where R indicates the relationship between the subject and earnings management activity.
3.3.3 Debt Contract Motives and Earnings Management
Prior research has also investigated the relationship between an increasing risk of
breaching debt covenants or lending contracts and earnings management activity.
Chen (2006) suggests that firms are more likely to be successful in loan or funding
applications where they have higher net incomes and gearing ratios that are well
below the industry accepted threshold of 50 per cent. Dechow et al (1996) state that a
primary determinant of earnings management is the desire to raise external financing
at a low cost and to avoid any debt covenant restrictions. In examining firms that have
violated lending contracts, DeFond and Jimbalvo (1994) determine that firms
accelerated their earnings, one year prior to the breach of debt covenants.
Additionally, Sweeney (1994) determines that debt covenant violators typically
engage in income increasing accounting policy changes; however this engagement is
generally post debt covenant violation.
Chapter 3 Literature Review
___________________________________________________________________________
22
3.3.4 Earnings Expectations and Earnings Management
Additional research has examined whether earnings management occurs in order to
meet or exceed the expectations of institutional investors, analysts and other capital
market participants. Payne and Robb (1997) determine that firms manage their
earnings in order to meet or comply with analysts’ forecasts. Kasznik (1999) provides
evidence consistent with the findings of Payne and Robb (1997), determining that
managers utilise positive discretionary accruals to inflate earnings when earnings
would otherwise, in the absence of inflation, fall below prior management forecasts.
Habib and Hansen (2008) state that the importance placed upon meeting analysts’
forecast benchmarks has increased in recent years. Lopez and Rees (2002) determine
via empirical analysis that 65 per cent of sample firms met or exceeded analysts’
forecasts during the years post 1992. Lopez and Rees (2002) also find that the
negative response of stock market participants to not meeting forecasts is significantly
greater, in absolute terms, than the response to beating forecasts and that meeting
analysts’ forecasts is a more powerful variable in the explanation of returns than the
annual profit or loss performance of a firm.
3.3.5 Regulatory Motives and Earnings Management
Regulatory motives refer to those within the context of either governmental regulation
(Healy and Wahlen, 1998) or self-regulation in the form of effective or defective
corporate governance mechanisms (Jouber and Fakhfakh, 2011). Jones (1991)
determines that U.S. firms seeking import duty relief generally depress earnings in the
year of application for such relief. Within the regulated U.S. banking sector and in
the context of the provision for credit loss, Healy and Wahlen (1998) state that there is
considerable evidence of excess loan loss provisioning and a subsequent
understatement of loan loss impairments to facilitate the recognition of abnormal
unrealised gains. This assertion is supported by Collins et al (1995) who determine
that profitable banks decrease their loan loss provisions when their earnings are
relatively low and increase such provisions when earnings are relatively high - this
being a clear indicator of pro-cyclical earnings management. Yun and Kim (2011)
infer that regulation can limit earnings management activity, finding that there has
been a significant decline in the proportion of discretionary accruals amongst sample
firms post implementation of the Sarbanes Oxley Act, 2002.
Chapter 3 Literature Review
___________________________________________________________________________
23
3.3.6 Corporate Governance Structures and Earnings Management
Empirical research has also identified a strong link between effective corporate
governance structures and a reduced level of earnings management activity,
as summarised in Table 3.2 below. Dechow et al (1996) find that poor oversight
through weak governance structures is an important determinant of earnings
manipulation. Chen (2006) determines that Taiwanese firms who are engaged in
earnings manipulation have a lower concentration of INEDs on their boards of
directors and supervisory boards, when compared with non-manipulating firms.
Peasnell et al (2005) posit that, as earnings management imposes costs upon
stockholders and capital market participants, effective corporate boards should work
towards preventing such manipulation, while also finding that the incidence of income
increasing earnings management activity decreases as the concentration of external
board members to the total board increases. Sebahattin and Harlan (2009) determine
that effective or strong corporate governance mechanisms within U.S. manufacturing
firms reduce the incidence of earnings management amongst mid-range firms. In light
of these findings, Sebahattin and Harlan (2009) also assert that creditors and equity
investors should apply greater scrutiny to the reported accruals of firms, being
mindful that robust corporate governance structures may represent an intervening
variable with regard to abnormal accruals.
Table 3.2 – Corporate Governance Structures and Earnings Management
Author Subject Earnings Management R
Dechow et al (1996) Weak Governance Increased Manipulation +
Beasley (1996) (+) INEDs to Board Reduction in Fraud -
Peasnell et al (2005) (+) INEDs to Board Reduced Earnings Inflation -
Sebahattin and Harlan (2009) Strong Governance Reduced Earnings Mgmt. -
Where R indicates the relationship between the subject and earnings management activity.
Additionally, Beasley (1996) determines that the proportion of external director
concentration to the board of directors for firms experiencing financial statement
fraud is lower when compared with non-fraud firms and that the composition of the
board, rather than audit committee presence, is a more important mechanism for
reducing the likelihood of financial statement manipulation.
Chapter 3 Literature Review
___________________________________________________________________________
24
3.3.7 Audit Committee and Earnings Management
The audit committee is a sub-component of the overall governance structure within a
firm. While Lin et al (2006) posit that there is a significantly negative association
between the independence of the audit committee and the incidence of earnings
restatement, along with a significantly negative association between the number of
audit committee meetings and the incidence of earnings restatement, neither of these
hypotheses is supported when subject to testing.
3.3.8 Auditor Type, Auditor Remuneration and Earnings Management
Francis and Krishnan (1999) determine that large audit firms provide higher quality
audits, exhibit reporting conservatism and are more aggressive in constraining the
earnings management activity of their clients. Krishnan (2003) determines that clients
of non-specialist auditors exhibit elevated levels of discretionary accruals when
compared with clients of specialist auditors while also suggesting that the use of a
Big 61 auditor with industry specialist knowledge can enhance the credibility of
accounting information.
Jordan et al (2010) determine that managers of companies audited by small audit
firms manipulate their earnings to generate EPS values consistent with user reference
points and that Big 4 audit firms are more likely to be effective in constraining the
efforts of their clients in earnings manipulation. In addition, Lin et al (2006) suggest
that higher audit fees from either audit specific or non-audit services reduce auditor
independence and therefore impair overall audit quality. In contrast, Frankel et al
(2002) find that there is a negative association between audit specific fees and
earnings management indicators. Table 3.3 below summarises these findings.
Table 3.3 – Auditor Type, Auditor Remuneration and Earnings Management
Author Subject Earnings Management R
Francis and Krishnan (1999) Large Audit Firms Reduced Earnings Mgmt. -
Frankel et al (2002) (+) Audit Fees Reduced Earnings Mgmt. -
Lin et al (2006) (+) Audit Fees Increased Earnings Mgmt. +
Jordan et al (2010) Small Audit Firms Increased Earnings Mgmt. +
Where R indicates the relationship between the subject and earnings management activity.
1 Big 6 audit firms in 2003 have subsequently reduced to Big 4 audit firms at the time of this study.
Chapter 3 Literature Review
___________________________________________________________________________
25
3.3.9 Revenue Manipulation, Deferred Revenue and Trade Receivables
The determinants of both revenue and trade receivables manipulation are discussed in
this section, given the importance of trade receivables throughout this study. Investors
and capital market participants place significant emphasis upon reported revenue
according to Anderson and Yohn (2002), who conclude that when there are
irregularities with a firm’s financial statements, investors are more concerned with
revenue recognition than alternative reporting issues, with revenue restatements
resulting in significantly more adverse stock returns compared with alternative
accounting restatements.
Caylor (2009) determines that managers engage in accelerated revenue recognition
using the short term deferred revenue and gross trade receivable accounts where
pre-managed earnings fall slightly below analyst benchmarks. Caylor (2009) also
determines that managers prefer to exercise revenue recognition in deferred revenue
rather than trade receivables in order to avoid negative earnings surprises.
Revenues form a unique role in valuations and it is preferable for managers to
manipulate revenues when compared with alternative earnings management methods,
as alternative earnings management methods are not equivalent in monetary outcomes
according to Zhang (2006); who also determines that firms with the following
characteristics are more likely to manage or manipulate revenues:
Higher growth perspectives
Higher operating margins
Outstanding analyst sales forecasts
Higher accounting policy flexibility in revenue recognition
Jeffrey et al (2008) determine that the greater a firm’s historical operating losses or
past and expected negative operating cash flows, the more likely it is to overstate
revenues and accounts receivable in order to induce a higher market valuation. Jeffrey
et al (2008) also determine that there is a positive relationship between the likelihood
of revenue manipulation and increasing leverage and inventory to total asset ratios.
Chapter 3 Literature Review
___________________________________________________________________________
26
3.4 The Response of Capital Markets
3.4.1 Earnings Quality and Stock Price Performance
While there is broad cross literature agreement that capital markets respond
negatively to low quality earnings, the available empirical evidence remains
somewhat conflicting. Dechow et al (2007) document a negative raw stock price
performance amongst manipulating firms during the years directly post earnings
manipulation. As indicated in Figure 3.1 below, there is a pronounced decline in stock
price performance post earnings manipulation, with only a slow recovery thereafter,
highlighting the longer term capital market effects arising from such activity.
Figure 3.1 – Stock Price Performance: Post Earnings Manipulation
Annual Raw Returns Surrounding the Earnings Manipulation Years
Year Relative to Manipulation Years
Note: Adapted directly from Dechow et al (2007).
Chan et al (2001) determine that there is a reliable negative association between
elevated levels of accruals, categorised as low quality earnings, and future stock
returns; also noting that changes in accounts receivable have strong predictive power.
However, Sloan (1996), in finding that the persistence of earnings performance is
dependent on the magnitude of the cash and accrual components of earnings,
determines that stock price results are inconsistent with the traditional efficient market
hypothesis that stock prices fully reflect all publicly available information.
Chapter 3 Literature Review
___________________________________________________________________________
27
3.4.2 Equity Offerings, Earnings Management and Stock Price Performance
Teoh et al (1998) determine that discretionary current accruals, which are subject to
managerial judgement, are artificially high around the period of an initial public
offering when compared with non-issuers. Teoh et al (1998) also determine that
issuers with an abnormally high level of discretionary accruals experience inferior
stock returns in the three years post initial public offering, with a firm in the most
aggressive category of initial public offering earnings managers experiencing, on
average, a 15 to 30 per cent worse three year stock price performance than those firms
classified as being within the most conservative range. This result is consistent with
the findings of Holthausen et al (1995) who determine that future stock returns are
negative for firms whose current earnings include large accrual components and
conversely, that future stock returns are positive for those firms with low accrual
components to their earnings.
3.4.3 Magnitude of Capital Market Based Earnings Management
While only limited prior research has measured the significance of capital market
based earnings management, Teoh et al (1994) determine, from a sample of firms
undertaking initial public offerings; that the median level of unexpected accruals
ranges from 4 – 5 per cent of total assets. Erickson and Wang (1999) find that
accruals for firms are measured at 2 per cent of total assets during the quarter of a
stock acquisition. While quantification of these findings in the absence of a
comparison with other non-issuing firms remains difficult, Teoh et al (1994)
determine that 62 per cent of firms undertaking initial public offerings exhibit
abnormally high levels of unexpected accruals when compared with a sample of
control firms, indicating an identifiable level of earnings management activity.
3.5 Conclusion
The numerous determinants of earnings management discussed throughout this
chapter provide a clear link between the previously cited accounting and economic
theory and extant earnings management practice, particularly with regard to the
opportunistic perspectives of positive accounting theory. Subsequently, varying
models have been developed to analyse the determinants of earnings management
activity. These models are considered in detail in the following chapter.
Chapter 4
Literature Review:
Testing For Earnings Management:
Model Development and Relevant
Accounting Standard Guidance
Chapter 4 Literature Review
___________________________________________________________________________
29
CHAPTER FOUR
LITERATURE REVIEW
Testing for Earnings Management: Model Development and
Relevant Accounting Standard Guidance
4.1 Introduction
The purpose of this chapter is to provide a chronological overview of the development
of models and measures utilised in testing for earnings quality and earnings
management, focusing on the provision for credit loss on trade receivables
specifically. Moreover, the chapter also details the current IFRS accounting standard
guidance and disclosure requirements with regard to the provision for credit loss on
trade receivables.
Chapter 4 Literature Review
___________________________________________________________________________
30
4.2 The Provision for Doubtful Receivables: Model Development
4.2.1 The Provision for Doubtful Receivables: McNichols and Wilson (1988)
There is a significant literature gap with regard to empirical studies that examine the
level of provision for credit loss on trade receivables, amongst IFRS compliant firms,
in the context of earnings management. This is possibly explained as a result of prior
research utilising a portfolio or combination approach in examining the combined
level of total discretionary accruals. In utilising the provision for doubtful debts as a
proxy for earnings management activity and predicting the provision for doubtful
debts in the absence of earnings management, McNichols and Wilson (1988)
determine that firms manage their earnings through the choice of income decreasing
accruals when income is extreme.
Additionally, McNichols and Wilson (1988) also determine that discretion in the
provision for bad debts ranges from 1 – 4 per cent of income for firms with extreme
income and that exercising discretion over the provision for bad debts, combined with
alternative discretionary accrual measures, can facilitate the achievement of target
income where annual earnings targets are within a 10 to 15 per cent growth range.
4.2.2 The Provision for Doubtful Receivables: Lev and Thiagarajan (1993)
In developing a multivariable earnings signal framework, Lev and Thiagarajan (1993)
identify disproportionate annual changes in trade receivables relative to revenue and
disproportionate annual changes in the provision for doubtful receivables relative to
trade receivables as fundamental indicators of earnings quality, suggesting that firms
with inadequate provisions for doubtful receivables are expected to experience future
depressed earnings as a result of increased provisions.
Lev and Thiagarajan (1993) also determine that while the aforementioned receivables
and doubtful receivables signals are relatively weak in unconditioned analysis, both
signals are statistically significant and value relevant during high inflation years,
indicating the importance of contextual or conditioned analysis. Lev and Thiagarajan
(1993) also suggest that there are adverse implications arising from inadequate bad
debt provisioning for both the persistence and growth of earnings.
Chapter 4 Literature Review
___________________________________________________________________________
31
4.2.3 Manipulation of Trade Receivables: Ricci (2011)
Ricci (2011) compares the receivables and receivables related accounts of companies
subject to U.S. SEC enforcement actions against those of a positive control group.
Utilising the Wilcoxon Signed Ranks Test, manipulating companies are paired against
non-manipulating companies within the same industry grouping. In determining that
trade receivables manipulation varies by industry type, Ricci (2011) finds that
receivables are inflated via the provision for doubtful receivables specifically; in the
Information Technology sector.
4.3 The Provision for Doubtful Receivables – IASB Accounting Guidance
4.3.1 IAS 39: Section 58 – 59
Section 58 of IAS 39 prescribes that an entity should reduce the carrying amount of a
financial asset either directly or through the use of an allowance account where
objective evidence of impairment exists at the reporting period end. Section 59 of IAS
39 also states that, inter alia, the following constitute objective evidence of
impairment:
“A breach of contract, such as a default or delinquency in interest or principal
payments”.
“National or local economic conditions that correlate with defaults on the assets
in a group”.
4.3.2 IFRS 7: Section 7.16 and 7.37
Section 7.16 of IFRS 7 stipulates that an entity must disclose a reconciliation of the
annual changes in the allowance account for each class of financial asset. Section 7.37
of IFRS 7 also stipulates that an entity must disclose:
“An analysis of the age of financial assets that are past due as at the end of the
reporting period but not impaired’’ and “An analysis of financial assets that are
individually determined to be impaired as at the end of the reporting period”.
These disclosures facilitate the determination of increasing or decreasing credit risk
arising on trade receivables from the examination of a firm’s financial statements.
Chapter 4 Literature Review
___________________________________________________________________________
32
4.4 Literature Review Conclusion
Capital market participants continue to place a significant emphasis upon the reported
earnings of firms (Schipper et al, 2003). As detailed in chapter two, the quality of
earnings is fundamental, as they represent a primary mechanism by which investors
can determine the most appropriate price of a security and attach a value to a firm.
However, there have been numerous corporate earnings scandals, arising from
earnings management practice, with resultant adverse effects for corporate
stakeholders (Bitner, 2005).
Accounting and economic theory provides strong support for the existence and
determinants of earnings management activity, while empirical research provides a
clear link between theory and extant earnings management practice. The majority of
these determinants have, however, been established in the context of earnings
management practice within the United States or U.S. GAAP compliant financial
reporting, as detailed in both chapters three and four. To this extent, there is a
significant literature gap with regard to empirical research in an IFRS compliant
financial reporting context. This gap is addressed in detail in the following chapter.
Chapter 5
Research
Methodology
Chapter 5 Research Methodology
___________________________________________________________________________
34
CHAPTER FIVE
RESEARCH METHODOLOGY
5.1 Introduction
The purpose of this chapter is to discuss the research rationale, research objectives
and research methodology underlying this study. The chapter firstly discusses the
research rationale for this study along with the research questions, before examining
the research objectives and literature driven research hypotheses. Thereafter, an
in-depth overview of the research methods undertaken in this study is provided, with
significant emphasis upon sample selection, data collection methods and the rationale
for the selection of various data measures. The chapter then concludes with an
overview of several limitations associated with this study.
Chapter 5 Research Methodology
___________________________________________________________________________
35
5.2 Research Rationale
Prior empirical studies, ranging from Healy (1985), Jones (1991) to Dechow et al
(2011) have examined in great detail the existence, frequency and magnitude of
earnings management activity, primarily in the context of discretionary accruals.
While Chen (2006) examines earnings management in an international context, the
majority of prior research has been conducted in a U.S. or U.S. GAAP compliant
financial reporting context. McNichols and Wilson (1988) identify the need for
further research with regard to earnings management through singular accrual
measures such as the provision for bad debts. However, excepting the research of Lev
and Thiagarajan (1993) and Ricci (2011), no extensive research has examined the
manipulation of the provision for credit loss on trade receivables in a European or
IFRS compliant financial reporting context.
Persistent macroeconomic uncertainty, particularly across Europe, combined with an
elevated level of credit risk relative to previous years (EFMA, 2012), renders this
study possible, timely and most appropriate. Undertaken in an IFRS compliant
financial reporting context, this study complements existing earnings management
research in adopting methodology employed in prior studies, while also adding to
existing literature. Moreover, this study provides a series of useful information for
various practitioners with regard to the magnitude and determinants of abnormal
provision for credit loss on trade receivables.
5.3 Research Questions
The research questions to be addressed in this study are:
RQ 1 – What is the magnitude and what are the determinants of abnormal provision
for credit loss on trade receivables amongst FTSE 350 companies?
RQ 2 – What is the capital (stock) market response to instances of extreme abnormal
provision for credit loss on trade receivables amongst FTSE 350 companies?
The research objectives that support the investigation of the research questions are
outlined overleaf.
Chapter 5 Research Methodology
___________________________________________________________________________
36
5.4 Research Objectives
Utilising abnormal change in the provision for credit loss on trade receivables as a
measure of earnings quality and as a proxy for earnings management activity:
1. To quantify the existence, direction and magnitude of abnormal provision for
credit loss on trade receivables amongst FTSE 350 companies.
2. To develop a multivariate OLS regression model that examines the
applicability of previously identified and alternative determinants of earnings
management activity, including capital market, contractual, performance,
governance and auditor related variables to abnormal provision for credit loss
on trade receivables amongst FTSE 350 companies.
3. To examine the individual and aggregate stock price performance of the most
extreme abnormal providers for credit loss on trade receivables (both under
and over providers) over a specified post financial year end period.
5.5 Research Hypotheses for Testing
This study, similar to previous earnings management studies, including Teoh et al
(1998), Frankel et al (2002) and Chen (2006) examines a large number of independent
explanatory variables. Research hypotheses relating to each research objective, along
with the associated independent variables and supporting literature references, where
applicable, are presented throughout.
5.5.1 – Objective One: Hypothesis
Maintaining the assertion that FTSE 350 companies are managing their earnings
through the provision for credit loss on trade receivables:
H1 – Ceteris paribus, the annual relative change in gross trade receivables does not
explain a significant extent of the variation in the annual relative change in the
provision for credit loss on trade receivables.
Literature Source: Lev and Thiagarajan (1993) – Lit. Review – Section 4.2.2
Chapter 5 Research Methodology
___________________________________________________________________________
37
5.5.2 – Objective Two: Hypotheses
The following hypotheses are conceptualised in the context of earnings inflation
activity, with abnormal underprovision for credit loss on trade receivables in an
environment of elevated credit risk representing earnings inflation activity.
Figure 5.1 – Capital Market Determinants of Earnings Management
H2 – Ceteris paribus, there is a negative association between analyst consensus EPS
growth forecasts and the abnormal change in provision for credit loss on trade
receivables.
H3 – Ceteris paribus, there is a negative association between a company’s earnings
(EPS) surprise and the abnormal change in provision for credit loss on trade
receivables.
Where: companies abnormally underprovide, to inflate earnings, in order to comply
with such capital market determinants.
Table 5.1 – Capital Market Determinants of Earnings Management
Literature Source: Literature Review: Section 3.3.4.
Payne and Robb (1997) - Kasznik (1999)
Lopez and Rees (2002) - Habib and Hansen (2008).
Figure 5.2 – Contractual Determinants of Earnings Management
Capital Market
Variables
Consensus EPS
Growth %
Earnings – EPS
Surprise %
Abnormal Change in
Provision for Credit
Loss on T-Receivables
Contractual
Variables
Existence of
Bonus Plan
Exec. Incentive
Remuneration
Abnormal Change in
Provision for Credit
Loss on T-Receivables
Change in
Gearing
Chapter 5 Research Methodology
___________________________________________________________________________
38
H4 – Ceteris paribus, there is a negative association between the existence of a bonus
plan and the abnormal change in provision for credit loss on trade receivables.
H5 – Ceteris paribus, there is a negative association between the proportion of
incentive (bonus) specific executive level remuneration and the abnormal change in
provision for credit loss on trade receivables.
H6 – Ceteris paribus, there is a negative association between the change in the level of
gearing of a firm and the abnormal change in provision for credit loss on trade
receivables.
Where: companies abnormally underprovide, to inflate earnings, given the existence
of such contractual incentives.
Table 5.2 – Contractual Determinants of Earnings Management
Literature Source: Literature Review: Sections 3.3.1 - 3.3.2 - 3.3.3.
Healy (1985) - Watts and Zimmerman (1986)
Dechow and Sloan (1991) - DeFond and Jimbalvo (1994)
Dechow et al (1996) - Balsam (1998) - Chen (2006).
Figure 5.3 – Performance Related Determinants of Earnings Management
H7 – Ceteris paribus, there is a negative association between the change in the gross
margin of a firm and the abnormal change in provision for credit loss on trade
receivables.
H8 – Ceteris paribus, there is a negative association between the change in the net
margin of a firm and the abnormal change in provision for credit loss on trade
receivables.
Performance
Variables
Change in
Gross Margin
Change in Net
Margin
Abnormal Change in
Provision for Credit
Loss on T-Receivables
Change in
T/Rec Days
Chapter 5 Research Methodology
___________________________________________________________________________
39
H9 – Ceteris paribus, there is a negative association between the change in the
average trade receivables collection period of a firm and the abnormal change in
provision for credit loss on trade receivables.
Where: companies abnormally underprovide, to inflate earnings, in order to maintain
a positive top line performance through to the final earnings performance or to detract
attention from an increasing average trade receivables collection period.
Table 5.3 – Performance Related Determinants of Earnings Management
Literature Source: Literature Review: Section 3.3.9.
Zhang (2006) - Jeffrey et al (2008).
Figure 5.4 – Governance Specific Determinants of Earnings Management
The final No. Of Gov. Non Compliance Issues variable has been introduced to
enhance the robustness of this study. Grant Thornton (2011) determines that only half
of all FTSE 350 companies were fully compliant with the Combined Code during
their 2011 review. As a result, the No. of Gov. Non Compliance Issues variable is
selected as a proxy for overall governance best practice within a FTSE 350 company.
Moreover, it is hypothesised that previously robust variables such as the proportion of
INEDs to the Audit Committee may no longer be robust, given the emergence of
corporate governance best practice in recent years.
Governance
Variables
INEDSs to
Total Board
INEDs to Audit
Committee
Abnormal Change in
Provision for Credit
Loss on T-Receivables
No. Of Audit
Committee
Meetings
No. Gov. Non
Compliance
Issues
Chapter 5 Research Methodology
___________________________________________________________________________
40
H10 – Ceteris paribus, there is a positive association between the proportion of INEDs
to the total board of directors and the abnormal change in provision for credit loss on
trade receivables.
H11 – Ceteris paribus, there is a positive association between the proportion of INEDs
to the total audit committee and the abnormal change in provision for credit loss on
trade receivables.
H12 – Ceteris paribus, there is a positive association between the number of audit
committee meetings held during the financial year and the abnormal change in
provision for credit loss on trade receivables.
H13 – Ceteris paribus, there is a negative association between the number of firm
specific governance non-compliance issues and the abnormal change in provision for
credit loss on trade receivables.
Where: robust corporate governance structures mitigate the propensity towards
earnings inflation activity, resulting in abnormal overprovision, categorised as prudent
activity throughout this study.
Table 5.4 – Governance Specific Determinants of Earnings Management
Literature Source: Literature Review: Sections 3.3.6 – 3.3.7.
Dechow et al (1996) - Beasley (1996)
Peasnell et al (2005) - Chen (2006) - Lin et al (2006).
Figure 5.5 – Auditor Related Determinants of Earnings Management
H14 – Ceteris paribus, there is a positive association between auditor type and the
abnormal change in provision for credit loss on trade receivables.
H15 – Ceteris paribus, there is a positive association between audit specific fees and
the abnormal change in provision for credit loss on trade receivables.
Auditor
Variables
Auditor Type
Audit Specific
Fee(s)
Abnormal Change in
Provision for Credit
Loss on T-Receivables
Chapter 5 Research Methodology
___________________________________________________________________________
41
Where: Big 4 auditors are more effective in constraining the abnormal underprovision
activity of their clients and where higher audit specific fees mitigate earning inflation
activity, resulting in abnormal overprovision.
Table 5.5 – Auditor Related Determinants of Earnings Management
Literature Source: Literature Review: Section 3.3.8.
Francis and Krishnan (1999) - Frankel et al (2002)
Krishnan (2003) - Lin et al (2006) - Jordan et al (2010)
5.5.3 – Objective Three: Hypothesis
In developing the following hypothesis, extreme abnormal underprovision for credit
loss on trade receivables is defined as earnings inflation activity. While extreme
abnormal overprovision may also be defined as earnings deflation activity, such
overprovision is categorised as being prudent rather than representing earnings
deflation activity in an environment of elevated credit risk.
H16 – Ceteris paribus, FTSE 350 companies with extreme abnormal underprovision
for credit loss on trade receivables (and as a result lower quality earnings) experience
an inferior stock price performance, post financial year end, relative to FTSE 350
companies with extreme abnormal overprovision for credit loss on trade receivables.
Table 5.6 – Capital Market Response to Earnings Management
Literature Source: Literature Review: Sections 3.4.1 – 3.4.3.
Holthausen et al (1995) - Teoh et al (1998)
Sloan (1996) - Chan et al (2001).
5.6 Research Approach
The research approach of this study is relatively consistent with prior earnings
management research, with the use of correlation and regression quantitative
techniques to analyse earnings management activity. Webster (1995, p.621) states that
while: “Regression determines if X and Y exhibit a positive relationship, or if the
relationship is negative in that they move in opposite directions, correlation measures
how strong the relationship is between X and Y” .
Chapter 5 Research Methodology
___________________________________________________________________________
42
The breadth of this study is, however, more extensive than prior research, in
considering the magnitude and determinants of earnings management activity, along
with the response of capital markets. Prior research has generally only examined two
of these facets simultaneously. However, such studies have often had access to
evidence of earnings management, primarily through the U.S. GAO accounting
restatement database (Jeffrey et al, 2008), eliminating the need to determine the
existence of earnings management.
5.6.1 Dependent Variable: Earnings Quality and Earnings Management
Where such evidence is not available, the proxy measure utilised for earnings quality
or evidence of earnings management generally comprises a measure of accounting
choice, including discretionary accruals or a measure of earnings irregularity relative
to analyst consensus expectations. Prior research, including Ricci (2011) has
examined firms specifically subject to U.S. SEC2 enforcement actions. As no earnings
restatement database is available in an IFRS compliant financial reporting context,
this study firstly determines the existence of abnormal provision for credit loss on
trade receivables, utilising this measure as an indicator of earnings quality and as a
proxy for earnings management activity. In conducting univariate and multivariate
regression analyses, the primary dependent variable comprises the relative change in
the provision for credit loss on trade receivables after controlling for the relative
change in gross trade receivables (Section 5.9.1) (Lev and Thiagarajan, 1993).
5.7 Sample Selection Process
5.7.1 Sample Selection Context
This study is undertaken to examine earnings quality and earnings management in a
European, IFRS compliant financial reporting context. As a result, the FTSE 350
Index, a market capitalisation weighted index incorporating all FTSE 100 and FTSE
250 companies is chosen as the initial sample population. This ensures consistency
with previous studies including Beasley (1996), Zhang (2006) and Jeffrey et al (2008)
which all contain large sample populations of greater than 100 firms.
2 U.S. SEC refers to the United States Securities and Exchange Commission.
Chapter 5 Research Methodology
___________________________________________________________________________
43
5.7.2 Sample Selection Refinement
The initial total sample of 350 companies relates to the composition of the FTSE 350
Index on 07 June 2012. Since this study has been undertaken, the researcher is aware
of only one company, Supergroup PLC; that has been demoted from the FTSE 350
Index. Company sector classification has been undertaken in accordance with the
FTSE 350 sector classifications of the London Stock Exchange (2012). Consistent
with prior studies, including Burgstahler and Dichev (1997), banks, financial and
financial related institutions are specifically excluded, given the significant variance
in their capital structures. Moreover, Collins et al (1995) already document the
existence of earnings management through abnormal provisioning activity within
banks, further supporting their exclusion from this study.
5.7.3 Sample Selection Refinement – Specific Elimination Procedures
Companies within sectors classified as Banks, Equity Investment Instruments,
Financial Services, Life Insurance, Non-Equity Investment Instruments, Nonlife
Insurance, Real Estate Investment and Services and Real Estate Investment Trusts
were therefore specifically excluded. Ruspetro PLC, admitted to the FTSE 350 Index
during early 2012, had not published any financial statements as at 07 June 2012 and
was therefore excluded from the study. After exclusion of these companies, the total
sample consisted of 241 companies.
A preliminary examination of their financial statements resulted in the exclusion of a
further 31 companies, due to the omission of information with regard to the provision
for credit loss on trade receivables or where summary receivables were presented, that
comprised significantly of VAT and other non-trade components. Of the remaining
210 companies, a further six were excluded where the extent of the relative change in
their provision for credit loss on trade receivables, after controlling for the relative
change in gross trade receivables, was extreme relative to the total sample population.
The characteristics of these six companies are discussed in the Research Findings
chapter. Full details surrounding the final sample population of 204 companies are
contained in Appendix C, while a summary of sample selection refinement is
contained in Table 5.7 overleaf.
Chapter 5 Research Methodology
___________________________________________________________________________
44
Table 5.7 – Sample Selection Refinement Summary
Total
Initial FTSE 350 Population 350
Exclusion of Financial Firms -108
Annual Report Unavailable -1
Insufficient Information – (Abnormal) -31
Outlier Firms -6
Final Sample 204
Information regarding the sector specific composition of the final sample of 204
companies is also contained in Appendix C.
5.8 Data Sources
Previous earnings management studies, including Zhang (2006) and Jeffrey et al
(2008) utilise the Compustat and CRSP3 databases for complete data collection. While
the Thomson One Banker database is used extensively throughout this study, in-depth
analysis of individual annual reports and disclosure notes is also necessary to collect
firm specific data relating to trade receivables, the provision for credit loss on trade
receivables, along with contractual, governance and auditor related variables.
The latest available annual reports for all companies were downloaded from the
Investor Relations sections of companies’ websites on 07 June 2012. A summary
analysis of the financial year end dates of the final sample of 204 companies is
contained in Table 5.8 below.
Table 5.8 – Financial Year End Dates Summary Analysis
Total
Financial Year End Date - 2010 1
Financial Year End Date - 2011 165
Financial Year End Date - 2012 38
Total Sample 204
3 CRSP refers to the Center for Research in Security Prices.
Chapter 5 Research Methodology
___________________________________________________________________________
45
While these financial year end dates are spread across a total of some 15 months
(31 December 2010 – 31 March 2012), the researcher deems this to be appropriate for
three reasons. Firstly, the datasets in previous studies, including Frankel et al (2002)
and Chen (2006) are multiannual in nature, with earnings management occurring in
alternating periods. Secondly, the exclusion of companies with a latest available
annual report for the year ended during 2011 or 2012 would have significantly
reduced the total sample size. Through the selection of the latest available annual
report for each firm, a degree of consistency is ensured. Thirdly, this study is
undertaken in an environment of elevated credit risk, which has remained elevated
post the 2008 financial crisis. Indeed, credit risk across Europe has further
deteriorated between 2011 and 2012 (EFMA, 2012). This inherent control ensures
that all data within the above range is gathered in the context of elevated credit risk.
Initial explanatory data was downloaded from the Thomson One Banker database on
15 June 2012. Data relating to the stock price performance of the 50 most extreme
abnormal providers for credit loss on trade receivables was downloaded on 26 June
2012. The source for data underlying each significant measure used in this study is
detailed in Table 5.9 below.
Table 5.9 – Data Sources for Data Measures
Data Measure Source Data Measure Source
Gross Trade Receivables AR Prop. Of INEDS to Total Board AR
Provision For Credit Loss On T/Rec AR Prop. Of INEDS to Audit Comm. AR
Consensus EPS Growth (%) TO # Of Gov. Non Compliance Issues AR
Earnings (EPS) Surprise (%) TO # Of Audit Committee Meetings AR
Existence of Bonus Plan AR Auditor Type AR
Exec. Incentive Remuneration AR Revenue AR
Change in Gross Margin (%) TO Audit Specific Fee(s) AR
Change in Net Margin (%) TO Stock Price Performance TO
Change in Gearing (%) TO Stock Beta Value TO
Change in Trade Rec. Days TO Total Assets TO
In Table 5.9 above, AR refers to the annual report of a company while TO refers to
the Thomson One Banker database.
Chapter 5 Research Methodology
___________________________________________________________________________
46
5.9 Data Measures Utilised
5.9.1 Primary Dependent Variable
As previously outlined, the primary dependent variable utilised in this study
comprises the relative change in the provision for credit loss on trade receivables after
controlling for the relative change in total gross trade receivables, which is calculated
as follows:
Measure of Abnormal Provision – Proxy for Earnings Management Activity
{ }-{ }: Where: EQ = Earnings quality.
Prov. (t) = Provision for credit loss on trade receivables in latest financial period.
Prov. (t-1) = Provision for credit loss on trade receivables in previous financial period.
GTR (t) = Gross trade receivables in latest financial period.
GTR (t-1) = Gross trade receivables in previous financial period.
Utilisation of this measure is directly consistent with Lev and Thiagarajan (1993).
However, this study reverses the direction of the measure to assist in classification of
the direction of abnormal provision, where underprovision results in a negative value
and overprovision results in a positive value. In doing so, this study does not deviate
from the underlying measure of Lev and Thiagarajan (1993), in measuring the relative
change in, and difference between, both variables.
Consistent with Lev and Thiagarajan (1993), this study attaches a single interpretation
to this measure, where abnormal underprovision is defined as earnings inflation
activity, in an environment of elevated credit risk. Throughout this study, this measure
is employed as the primary measure of earnings quality and proxy for earnings
management activity.
Prov. (t) - Prov. (t-1)
Prov. (t-1)
X 100
GTR (t) - GTR (t-1)
GTR (t-1)
X 100 EQ
Chapter 5 Research Methodology
___________________________________________________________________________
47
5.9.2 Additional Variables and Measures
The following is an overview of the dependent variable utilised in determining
whether the extent of abnormal provision for credit loss on trade receivables
significantly explains subsequent stock price performance. Full explanatory detail
relating to all additional measures employed in this study is contained in Appendix G.
Stock Price Performance %
{ } Where: Price Close (t) = Stock price close on 07
June 2012.
Price Close (t-1) = Stock price close on date of latest available financial year end.
While this stock price performance measure results in a variation in the period of
stock price performance analysis, it ensures a degree of consistency where the
response of capital markets to the latest available annual financial information of a
firm is considered. The 07 June 2012 is selected as the performance period cut-off
point, given that the study sample population was gathered on this date. This approach
is consistent with Dechow et al (2007), in analysing the raw stock price performance
of companies surrounding earnings manipulation years.
5.10 Data Validity and Reliability
Howell (2009, p.53) states that outliers deserve special attention as they may represent
erroneous measurement or recording procedures. While the six outliers identified in
this study were confirmed as being legitimate values, they were specifically excluded
and subject to individual analysis, in order to maintain the assumption of normal data
distribution and to enhance the reliability of regression analyses. The final dataset
relating to the magnitude and determinants of earnings management was complete,
without any missing variables. However, stock price performance data was
unavailable for two companies, therefore the subsequent most extreme abnormal
(under and over) providers for credit loss on trade receivables replaced these
companies, to ensure that the stock price performance sample population consisted of
an equal number of extreme abnormal under and over providers.
Price Close (t) – Price Close (t-1)
Price Close (t -1)
X 100
Chapter 5 Research Methodology
___________________________________________________________________________
48
5.11 Testing Procedures
The Microsoft Excel 2010 Data Analysis Toolpak is utilised to undertake univariate
and multivariate analyses to examine H1 through H16 inclusive. Consistent with Chen
(2006), multivariate OLS regression analyses comprises the majority of statistical
analysis undertaken, while descriptive statistics relating to each variable are also
generated. Although the multivariate analyses conducted is consistent with best
practice, it is subject to both inherent limitations and assumptions (Webster, 1995,
p.638). Extensive testing procedures that ensure compliance with these assumptions
are contained in Appendix G.
5.12 Limitations
In utilising abnormal provision for credit loss on trade receivables as the sole
indicator of earnings quality and proxy for earnings management activity, it is
possible that instances of earnings management through alternative abnormal
provisioning or discretionary accruals remain undetected. Equally, while companies
may have engaged in significant abnormal provision for credit loss on trade
receivables, they may not be engaged in earnings management across other variables,
with earnings management therefore insignificant in an overall context. While a single
earnings inflation motive is attached to abnormal underprovision for credit loss on
trade receivables, consistent with previous research, this approach disregards the
possibility that abnormal overprovision may represent earnings deflation activity.
Additionally, underprovision may well be justified, where the credit risk attaching to
specific customers has declined significantly.
These limitations are, however, considerably moderated, given that this study is
undertaken in an environment of elevated credit risk, rendering any underprovision
suspect. Moreover, given the pressure on firm specific revenues in a difficult
macroeconomic environment combined with capital market earnings expectations, it
is more unlikely that companies would engage in earnings deflation activity at the
present time. This study also assumes that the prior year provision for credit loss on
trade receivables is representative of steady state, un-managed, normal provisioning.
It is possible that provisioning activity was elevated during 2008 and 2009, amidst the
core of the financial crisis, with companies now reducing their provision for credit
loss on trade receivables in subsequent years.
Chapter 5 Research Methodology
___________________________________________________________________________
49
5.13 Conclusion
This chapter has provided an overview of the methodology underlying this study,
along with extensive detail regarding the research questions, objectives and
hypotheses that are subject to testing. The rationale for the sample selection
refinement procedures adopted and data measures selected has been discussed,
providing a firm basis for the research findings discussed in the following chapter.
Chapter 6
Research Findings
“Competency in mathematics, both in numerical manipulations and in
understanding its conceptual foundations, enhances a person’s ability to
handle the more ambiguous and qualitative relationships that dominate
our day-to-day financial decision-making”.
Alan Greenspan (2007)
Chapter 6 Research Findings
___________________________________________________________________________
51
CHAPTER SIX
RESEARCH FINDINGS
6.1 Introduction
The findings of the statistical analyses undertaken are presented sequentially in this
chapter, consistent with the order of the research objectives. Starting with an analysis
of the identified outliers; results relating to the three primary research objectives are
then presented and discussed. The discussion and presentation of findings relating to
each research objective begins with a presentation of descriptive statistics, followed
with the results of univariate and multivariate regression analyses before concluding
with acceptance or rejection of the research hypotheses.
Chapter 6 Research Findings
___________________________________________________________________________
52
6.2 Analysis of Identified Outliers
Outliers in this study are identified with reference to the primary dependent variable.
Six companies, summarised in Table 6.1 below, were identified as having extreme
abnormal provision for credit loss on trade receivables relative to the total sample,
which resulted in their exclusion from all subsequent analysis and reduced the final
sample population to 204 companies.
Table 6.1 – Six Identified Outliers
Company Name Abnormal Provision For Credit Loss
HERITAGE OIL PLC -2262.5%
SOCO INTERNATIONAL PLC -821.9%
KENMARE RESOURCES PLC -294.1%
FILTRONA PLC 156.3%
RANK GROUP PLC 181.0%
INTL CON. AIRLINES GROUP PLC 199.4%
The extent of abnormal underprovision was considerably more extreme than the
extent of abnormal overprovision amongst these six outliers. 33 per cent of outliers
comprised companies in the Oil and Gas sector - Heritage Oil PLC and SOCO
International PLC. A further 33 per cent comprised companies in the Travel and
Leisure sector - Intl Consolidated Airlines Group PLC and Rank Group PLC.
However, the moderately excessive overprovision in the case of Intl Consolidated
Airlines Group PLC arose through business combination activities. The remaining
two outliers - Kenmare Resources PLC and Filtrona PLC, comprised companies in the
Mining and Support Services sectors respectively. All six outliers were subsequently
excluded from further analysis, while Appendix D details the impact of these outliers
upon preliminary testing.
6.3 Magnitude of Abnormal Provision for Credit Loss on Trade Receivables
6.3.1 Descriptive Statistics
During the latest period, the mean relative increase in gross trade receivables is
measured at 12.0 per cent, with a corresponding mean relative increase in the
provision for credit loss on trade receivables of only 2.1 per cent. The mean level of
abnormal provision is therefore measured at -9.9 per cent (Table 6.2 overleaf).
Chapter 6 Research Findings
___________________________________________________________________________
53
Table 6.2 – Descriptive Statistics for Magnitude of Abnormal Provision
Magnitude of Abnormal Provision Mean Std
Dev Min Max Skew
Abnormal Provision for Credit Loss (%)* -9.90 0.40 -141.60 135.30 -0.10
Change in Gross Trade Receivables (%)* 12.00 0.30 -85.20 163.56 1.80
Change in Provision for Credit Loss (%)* 2.10 0.40 -73.28 191.20 2.10
All Percentages (Indicated*) are converted to Percentage Point Scores through Multiplication – (X 100)
The mean level of abnormal provision indicates that there was an average
underprovision for credit loss on trade receivables of 9.9 per cent, with a standard
deviation value of 0.4 suggesting that the majority of values lie close to this mean.
Of the total sample of 204 FTSE 350 companies, 138 or 67.7 per cent exhibited
abnormal underprovision during the period. Consistent with the range of abnormal
provision for credit loss on trade receivables, both the relative change in gross trade
receivables and the relative change in the provision for credit loss on trade receivables
exhibit wide ranges of -85.2 per cent to 163.6 per cent and -73.28 per cent to 191.2
per cent respectively. The skewness value of -0.1 also indicates that the abnormal
provision for credit loss on trade receivables variable is normally distributed.
6.3.2 Univariate Analysis: Simple Linear Regression Analysis
OLS simple regression analysis is utilised to determine the extent to which the
relative change in gross trade receivables explains the relative change in the provision
for credit loss on trade receivables, as denoted in Figure 6.1 below.
Figure 6.1 – Simple Linear Regression Equation
Table 6.3 - Simple Linear Regression Significance Statistics
β0 Constant R Square 0.103
Coefficient -0.020 Adj. R Square 0.099
T-Stat -0.816 F Stat 23.229
P-Value 0.415 P-Value F Stat 0.000
∆% in Prov. for Credit Loss on T/Rec = β0 + β1∆% in Gross Trade Receivables + 𝜀
Chapter 6 Research Findings
___________________________________________________________________________
54
Table 6.3 - Simple Linear Regression Significance Statistics Continued
Detail β1 ∆% in GTR
Predicted Sign +
Coefficient 0.340
T-Stat 4.820
P-Value 0.000
As anticipated, the coefficient of β1 (0.340) is positive, with a P-Value of 0.000
indicating significance at the 1% level, confirming that there is a significant positive
relationship between the relative change in gross trade receivables and the relative
change in the provision for credit loss on trade receivables. The regression
significance P-Value of 0.000 indicates that the model has explanatory power at all
levels of significance. While both variables exhibit a positive association, the Adj. R
Square value of 0.099 indicates that only 9.9 per cent of the variation in the relative
change in provision for credit loss on trade receivables is explained by the relative
change in gross trade receivables. H1 is therefore accepted.
6.4 Determinants of Abnormal Provision for Credit Loss on Trade Receivables
6.4.1 Compliance with Underlying OLS Regression Analysis Assumptions
The utilisation of univariate and multivariate regression analyses requires compliance
with the underlying assumptions of OLS regression analysis, including correct model
specification, normal distribution of the error observations (𝜀i) and the absence of
multicollinearity between the independent explanatory variables. As the results of all
procedures contained in Appendix G verify compliance with these underlying
assumptions, the researcher deems all findings relating to both univariate and
multivariate analyses conducted to be reliable.
6.4.2 Descriptive Statistics: Independent Explanatory Variables
Explanatory detail relating to all independent variables is detailed in Appendix G.
Of the 14 independent explanatory variables utilised in multivariate regression
analysis, 12 are continuous in nature, while two are categorical in nature (Table 6.4
overleaf).
Chapter 6 Research Findings
___________________________________________________________________________
55
Table 6.4 – Descriptive Statistics for Categorical Independent Variables
FREQ 1 FREQ 0
β3 Existence of Bonus Plan 202 2
β13 Auditor Type 198 6
As indicated in Table 6.4 above, 202 or 99 per cent of the 204 FTSE 350 companies
operated bonus or incentive remuneration related plans during the period, while 198
or 97 per cent of the 204 FTSE 350 companies were audited by a Big 4 auditor.
Table 6.5 – Descriptive Statistics for Continuous Independent Variables
Determinants of Earnings Mgmt. Mean Std
Dev Min Max Skew
β1 Consensus EPS Growth (%)* 34.91 1.56 -63.48 1760.00 8.72
β2 Earnings (EPS) Surprise (%)* 0.33 0.21 -185.00 97.00 -4.24
β4 Executive Incentive Remuneration 0.36 0.18 0.00 0.89 -0.41
β5 Change in Gearing (%)* 0.09 5.77 -23.44 21.00 0.21
β6 Change in Gross Margin (%)* 1.03 6.63 -18.39 51.89 3.73
β7 Change in Net Margin (%)* 0.04 10.31 -72.95 41.28 -2.20
β8 Change in Trade Rec. Days 0.88 20.41 -74.93 239.65 7.66
β9 Proportion of INEDs to the Total Board* 53.95 0.10 28.57 80.00 0.08
β10 Proportion of INEDs to Audit Comm.* 98.00 0.07 50.00 100.00 -5.48
β11 Of Governance Non Compliance Issues 0.74 1.09 0.00 8.00 2.68
β12 Of Audit Committee Meetings 4.32 1.58 2.00 14.00 2.39
β14 Audit Specific Fee(s) 0.08 0.06 0.01 0.44 1.85
All Percentages (Indicated*) are converted to Percentage Point Scores through Multiplication – (X 100)
Of the 12 continuous independent variables, both the Earnings (EPS) Surprise (%)
and Proportion of INEDs to Audit Committee variables are significantly negatively
skewed (Table 6.5 above), while the Consensus EPS Growth (%) variable exhibits
significantly positive skew, indicating abnormal distribution in these instances.
The skewness values of the Executive Incentive Remuneration, Change in Gearing
and Proportion of INEDs to the Total Board variables indicate normal distribution.
Chapter 6 Research Findings
___________________________________________________________________________
56
The mean of the Consensus EPS Growth (%) variable (34.91%) indicates that analyst
expectations were for significant EPS and earnings growth during the latest period,
with an extreme EPS growth rate expectation of 1760% in one instance. The mean
values of the Proportion of INEDs to the Total Board (53.95%), Proportion of INEDs
to Audit Committee (98.00%) and No. of Governance Non Compliance Issues (0.74)
variables indicate generally strong compliance with corporate governance best
practice and the Combined Code amongst the 204 FTSE 350 companies.
Finally, the mean increase of 0.88 days in the average trade receivables collection
period (Change in Trade Rec. Days) indicates that there was very limited extension of
trade credit during the period amongst the 204 FTSE 350 companies.
6.4.3 Univariate Analysis: Simple Linear Regression Analysis
OLS simple regression analysis was conducted to examine the extent of univariate
relationships between the primary dependent variable: Earnings Quality (proxy for
earnings management activity) and each of the previously identified 14 independent
variables. Initially undertaken with the full sample of 204 companies, simple
regression analysis was subsequently restricted to the 138 identified underproviders,
as the hypotheses in this study are conceptualised in the context of earnings inflation
activity.
Figure 6.2 – Simple Linear Regression Equation
Where:
Earnings Quality = Abnormal provision for credit loss on trade receivables and proxy
for earnings management. βXi = Each of the 14 independent explanatory variables.
β0 = Intercept and 𝜀 = Regression error term.
The results of simple linear regression analysis are contained in Table 6.6 overleaf.
Earnings Quality = β0 + βXi + 𝜀
Chapter 6 Research Findings
___________________________________________________________________________
57
Table 6.6 – Simple Regression Analysis Results: Full Sample (N=204)
Hypo. Variable Pred.
Sign Coefficient P-Value
2 β1 Consensus EPS Growth (%) - -0.023 0.181
3 β2 Earnings (EPS) Surprise (%) - 0.188 0.151
4 β3 Existence of Bonus Plan - 0.046 0.866
5 β4 Executive Incentive Remuneration - -0.048 0.866
6 β5 Change in Gearing (%) - -0.011 0.022
7 β6 Change in Gross Margin (%) - 0.004 0.369
8 β7 Change in Net Margin (%) - 0.003 0.203
9 β8 Change in Trade Rec. Days - -0.001 0.669
10 β9 Proportion of INEDs to the Total Board + 0.543 0.036
11 β10 Proportion of INEDs to Audit Committee + 0.246 0.524
12 β11 # Of Governance Non Compliance Issues - -0.008 0.748
13 β12 # Of Audit Committee Meetings + 0.008 0.655
14 β13 Auditor Type + 0.010 0.524
15 β14 Audit Specific Fee(s) + -0.270 0.555
Bolded P-Values in the above Table indicate statistical significance at the 5% level.
As indicated in Table 6.6 above, 12 of the 14 independent variables exhibit no
significant influence on the direction of abnormal provision for credit loss on trade
receivables. However, the bolded P-Values, significant at the 5% level, indicate that
both the Change in Gearing and Proportion of INEDs to the Total Board variables
have a significant influence on the direction of abnormal provision for credit loss on
trade receivables, with the signs of the coefficients consistent with those hypothesised
in both instances.
Table 6.7 – Simple Regression Analysis Results: Underproviders Only (N=138)
Hypo. Variable Pred.
Sign Coefficient P-Value
2 β1 Consensus EPS Growth (%) - -0.015 0.266
3 β2 Earnings (EPS) Surprise (%) - 0.156 0.161
4 β3 Existence of Bonus Plan - -0.134 0.508
5 β4 Executive Incentive Remuneration - 0.038 0.485
6 β5 Change in Gearing (%) - -0.007 0.100
7 β6 Change in Gross Margin (%) - -0.009 0.039
Chapter 6 Research Findings
___________________________________________________________________________
58
Table 6.7 – Simple Regression Analysis: Underproviders Only (N=138)
Hypo. Variable Pred.
Sign Coefficient P-Value
8 β7 Change in Net Margin (%) - 0.002 0.347
9 β8 Change in Trade Rec. Days - 0.001 0.587
10 β9 Proportion of INEDs to the Total Board + 0.483 0.037
11 β10 Proportion of INEDs to Audit Committee + -0.143 0.641
12 β11 # Of Governance Non Compliance Issues - -0.001 0.981
13 β12 # Of Audit Committee Meetings + 0.008 0.627
14 β13 Auditor Type + 0.377 0.022*
15 β14 Audit Specific Fee(s) + -0.189 0.628
Bolded P-Values in the above Table indicate statistical significance at the 5% or 10% level.
Table 6.7 above indicates that amongst the restricted sample of 138 underproviders,
the Change in Gross Margin and Proportion of INEDs to the Total Board variables
significantly influence the direction of abnormal provision for credit loss on trade
receivables at the 5% level. The Change in Gearing also exhibits significance at the
10% level. The signs of the coefficients of all significant variables are also consistent
with those hypothesised. While the Auditor Type variable exhibits an apparent
significant relationship (0.022*), this result contrasts sharply with the previous
univariate analysis and is both deemed to be skewed and disregarded, given that this
variable is not normally distributed, where only three of the 138 abnormal
underproviders were audited by a non-Big 4 auditor.
6.4.4 Multivariate Analysis: Multiple Regression Analysis
OLS multiple regression analysis was conducted to examine the extent of multivariate
relationships between the primary dependent variable: Earnings Quality (proxy for
earnings management activity) and each of the previously identified 14 independent
variables. The analysis was also further refined with the exclusion of insignificant
variables, restriction of the analysis to the 138 identified underproviders, along with
multiple regression analysis undertaken with the secondary dependent variable as
detailed in Appendix D.
Regression One: Fourteen Hypotheses Regression (N=204)
Regression one consists of the full sample of 204 FTSE 350 companies and tests all
fourteen hypotheses as indicated in Figure 6.3 overleaf.
Chapter 6 Research Findings
___________________________________________________________________________
59
Figure 6.3 – Multiple Regression One Equation
The title of each variable in the equation is shortened in the interest of brevity.
Table 6.8 - Multiple Regression One Significance Statistics (N=204)
β0 Constant R Square 0.083
Coefficient -0.538 Adj. R Square 0.016
T-Stat -0.921 F Stat 1.215
P-Value 0.358 P-Value F Stat 0.266
The results contained in Table 6.8 above indicate limited explanatory power, with a
regression significance P-Value of 0.266 indicating insignificance at all levels. The
Adj. R Square value of 0.016 indicates that the model only explains 1.6 per cent of the
variation in the direction of abnormal provision for credit loss on trade receivables.
The results for all 14 hypotheses tested are contained in Table 6.9 below-overleaf.
Table 6.9 - Multiple Regression One: Fourteen Hypotheses (N=204)
Hypo. Variable Pred.
Sign Coefficient P-Value
2 β1 Consensus EPS Growth (%) - -0.025 0.173
3 β2 Earnings (EPS) Surprise (%) - 0.197 0.173
4 β3 Existence of Bonus Plan - 0.010 0.971
5 β4 Executive Incentive Remuneration - -0.075 0.296
6 β5 Change in Gearing (%) - -0.008 0.100
7 β6 Change in Gross Margin (%) - -0.005 0.296
8 β7 Change in Net Margin (%) - 0.005 0.091
Bolded P-Values in the above Table indicate statistical significance at the 5% or 10% level.
Earnings Quality = β0 + β1 EPS GROWTH+ β2 EPS SURPRISE + β3 BONUS
+ β4 REMUN + β5 GEARING + β6 GROSS MARGIN
+ β7 NET MARGIN + β8 T/REC DAYS + β9 INEDS BOARD
+ β10 INEDS AUDIT + β11 GOV NON COMP
+ β12 AUDIT MEET + β13 AUDITOR + β14 AUDIT FEES + 𝜀
Chapter 6 Research Findings
___________________________________________________________________________
60
Table 6.9 - Multiple Regression One: Fourteen Hypotheses Continued
Hypo. Variable Pred.
Sign Coefficient P-Value
9 β8 Change in Trade Rec. Days - -0.001 0.589
10 β9 Proportion of INEDs to the Total Board + 0.565 0.045
11 β10 Proportion of INEDs to Audit Committee + 0.206 0.645
12 β11 # Of Governance Non Compliance Issues - 0.014 0.646
13 β12 # Of Audit Committee Meetings + -0.001 0.960
14 β13 Auditor Type + -0.061 0.709
15 β14 Audit Specific Fee(s) + 0.221 0.650
Bolded P-Values in the above Table indicate statistical significance at the 5% or 10% level.
As outlined in Table 6.9, only three variables exhibit a statistically significant
relationship with the direction of abnormal provision for credit loss on trade
receivables: the Change in Net Margin (10% level), Change in Gearing (10% level),
and Proportion of INEDs to the Total Board (5% level). Moreover, the positive
coefficient of the Change in Net Margin variable is contrary to that hypothesised,
while the P-Values of several variables highlight their insignificance.
Regression Two: Seven Hypotheses Regression (N=204)
Regression two consists of the full sample of 204 FTSE 350 companies and tests
seven hypotheses, having eliminated the most insignificant variables, as indicated in
Figure 6.4 below.
Figure 6.4 – Multiple Regression Two Equation
The title of each variable in the equation is shortened in the interest of brevity.
The results contained in Table 6.10 overleaf indicate a marked improvement in the
explanatory power of the model, with a regression significance P-Value of 0.023 and
an Adj. R Square value of 0.046 indicating that the model explains 4.6 per cent of the
variation in the direction of abnormal provision for credit loss on trade receivables.
Earnings Quality = β0 + β1 EPS GROWTH + β2 EPS SURPRISE + β4 REMUN
+ β5 GEARING + β6 GROSS MARGIN + β7 NET MARGIN
+ β9 INEDS BOARD + 𝜀
Chapter 6 Research Findings
___________________________________________________________________________
61
Table 6.10 - Multiple Regression Two Significance Statistics (N=204)
β0 Constant R Square 0.078
Coefficient -0.349 Adj. R Square 0.046
T-Stat -2.436 F Stat 2.379
P-Value 0.015 P-Value F Stat 0.023
Table 6.11 - Multiple Regression Two: Seven Hypotheses (N=204)
Hypo. Variable Pred.
Sign Coefficient P-Value
2 β1 Consensus EPS Growth (%) - -0.023 0.193
3 β2 Earnings (EPS) Surprise (%) - 0.187 0.167
5 β4 Executive Incentive Remuneration - -0.072 0.296
6 β5 Change in Gearing (%) - -0.008 0.084
7 β6 Change in Gross Margin (%) - -0.004 0.304
8 β7 Change in Net Margin (%) - 0.004 0.119
10 β9 Proportion of INEDs to the Total Board + 0.540 0.039
Bolded P-Values in the above Table indicate statistical significance at the 5% or 10% level.
Despite the marked improvement in the explanatory power of the model, only two
variables exhibit statistical significance in determining the direction of abnormal
provision for credit loss on trade receivables: the Change in Gearing (10% level) and
Proportion of INEDs to the Total Board (5% level). Moreover, the signs of the
coefficients of both variables are consistent with those hypothesised.
Regression Three: Fourteen Hypotheses Regression (N=138)
Regression three consists of the restricted sample of 138 underproviders and tests all
fourteen hypotheses as indicated in Figure 6.5 below.
Figure 6.5 – Multiple Regression Three Equation
The title of each variable in the equation is shortened in the interest of brevity.
Earnings Quality = β0 + β1 EPS GROWTH+ β2 EPS SURPRISE + β3 BONUS
+ β4 REMUN + β5 GEARING + β6 GROSS MARGIN
+ β7 NET MARGIN + β8 T/REC DAYS + β9 INEDS BOARD
+ β10 INEDS AUDIT + β11 GOV NON COMP
+ β12 AUDIT MEET + β13 AUDITOR + β14 AUDIT FEES + 𝜀
Chapter 6 Research Findings
___________________________________________________________________________
62
Table 6.12 - Multiple Regression Three Significance Statistics (N=138)
β0 Constant R Square 0.153
Coefficient -0.359 Adj. R Square 0.056
T-Stat -0.754 F Stat 1.583
P-Value 0.4522 P-Value F Stat 0.093
The Adj. R. Square value of 0.056 indicates that the fourteen hypotheses model, when
restricted to the 138 identified underproviders, explains 5.6 per cent of the variation in
the direction of abnormal provision for credit loss on trade receivables. While the
model still retains overall explanatory significance at the 10% level (P-Value 0.093),
a marked reduction in the F Stat is noted relative to regression two.
Table 6.13 - Multiple Regression Three: Fourteen Hypotheses (N=138)
Hypo. Variable Pred.
Sign Coefficient P-Value
2 β1 Consensus EPS Growth (%) - -0.011 0.400
3 β2 Earnings (EPS) Surprise (%) - 0.146 0.231
4 β3 Existence of Bonus Plan - -0.203 0.328
5 β4 Executive Incentive Remuneration - 0.022 0.688
6 β5 Change in Gearing (%) - -0.005 0.268
7 β6 Change in Gross Margin (%) - -0.009 0.050
8 β7 Change in Net Margin (%) - 0.001 0.530
9 β8 Change in Trade Rec. Days - 0.001 0.554
10 β9 Proportion of INEDs to the Total Board + 0.535 0.031
11 β10 Proportion of INEDs to Audit Committee + -0.335 0.360
12 β11 # Of Governance Non Compliance Issues - -0.003 0.911
13 β12 # Of Audit Committee Meetings + 0.004 0.819
14 β13 Auditor Type + 0.328 0.060*
15 β14 Audit Specific Fee(s) + 0.014 0.973
Bolded P-Values in the above Table indicate statistical significance at the 5% level.
As indicated in Table 6.13 above, the Change in Gross Margin (5% level) and
Proportion of INEDs to the Total Board (5% level) variables exhibit a significant
relationship with the direction of abnormal provision for credit loss on trade
receivables. However, the apparent significance of the Auditor Type variable (0.060*)
is deemed to be skewed and is disregarded (Refer to Table 6.7).
Chapter 6 Research Findings
___________________________________________________________________________
63
Regression Four: Four Hypotheses Regression (N=138)
Regression four consists of the restricted sample of 138 underproviders and tests four
hypotheses, having eliminated the most insignificant variables, as indicated in Figure
6.6 below.
Figure 6.6 – Multiple Regression Four Equation
The title of each variable in the equation is shortened in the interest of brevity.
Table 6.14 - Multiple Regression Four Significance Statistics (N=138)
β0 Constant R Square 0.116
Coefficient -0.867 Adj. R Square 0.089
T-Stat -4.489 F Stat 4.370
P-Value 0.000 P-Value F Stat 0.002
The significance statistics of regression four are the most robust of all four multiple
regressions. With an Adj. R Square value of 0.089 and a regression significance
P-Value of 0.002, the regression has explanatory power at all levels of significance,
explaining 8.9 per cent of the variation in the direction of abnormal provision for
credit loss on trade receivables, which is to be anticipated given the removal of
insignificant variables.
Table 6.15 – Multiple Regression Four: Four Hypotheses (N=138)
Hypo. Variable Pred.
Sign Coefficient P-Value
6 β5 Change in Gearing (%) - -0.005 0.201
7 β6 Change in Gross Margin (%) - -0.010 0.019
10 β9 Proportion of INEDs to the Total Board + 0.516 0.023
14 β13 Auditor Type + 0.334 0.038*
Bolded P-Values in the above Table indicate statistical significance at the 5% level.
As indicated in Table 6.15 above, the signs of the coefficients of all four independent
variables are consistent with those hypothesised.
Earnings Quality = β0 + β5 GEARING + β6 GROSS MARGIN + β9 INEDS BOARD
+ β13 AUDITOR + 𝜀
Chapter 6 Research Findings
___________________________________________________________________________
64
Indeed, two of the four variables: the Change in Gross Margin and Proportion of
INEDs to the Total Board exhibit a significant relationship with the direction of
abnormal provision for credit loss on trade receivables. The apparent significance of
the Auditor Type variable (0.038*) is once again disregarded (Refer to Table 6.7).
6.4.5 H2 to H15: Summary Findings
Acceptance or rejection of the research hypotheses (H2 – H15 inclusive) is undertaken
relative to the results of univariate regression analysis and multiple regressions two
through four inclusive, with the results of multiple regression one excluded given its
insignificant explanatory power.
H2 – H15 are conceptualised with regard to the abnormal change in provision for
credit loss on trade receivables. The dependent variable utilised in the aforementioned
analysis is the prime indicator of such change and also captures the extent of its
abnormality. The results of the analysis, contained in Table 6.16, indicate that:
Table 6.16 – H2 to H15: Summary Findings
There is a significant negative association between the Change in Gearing and the
direction of abnormal provision for credit loss on trade receivables. H6 is
therefore accepted.
There is a significant negative association between the Change in Gross Margin
and the direction of abnormal provision for credit loss on trade receivables. H7 is
therefore accepted.
There is a significant positive association between the Proportion of INEDs to the
Total Board and the direction of abnormal provision for credit loss on trade
receivables. H10 is therefore accepted.
In light of these findings, all additional research hypotheses H2 to H15 not detailed in
Table 6.16 are therefore rejected.
Chapter 6 Research Findings
___________________________________________________________________________
65
6.5 Stock Price Performance of Extreme Abnormal Providers
6.5.1 Descriptive Statistics
Descriptive statistics relating to the 25 most extreme abnormal underproviders for
credit loss on trade receivables are firstly presented in Table 6.17 below, while
descriptive statistics relating to the 25 most extreme abnormal overproviders are
presented in Table 6.18 thereafter.
Table 6.17 – Descriptive Statistics for Stock Price Performance: Underproviders
Extreme Abnormal Underproviders Mean Std
Dev Min Max Skew
Raw Stock Price Performance (%)* -11.10 0.28 -81.41 23.55 -0.87
Stock Beta: Risk Variable 0.97 0.60 -0.28 2.38 0.16
Natural Log of Total Assets: Size Variable 2.94 0.51 1.89 3.83 -0.26
All Percentages (Indicated*) are converted to Percentage Point Scores through Multiplication – (X 100)
As detailed in Table 6.17 above, the 25 most extreme abnormal underproviders
experienced an average post financial year end stock price performance of -11.1%,
with a standard deviation value of 0.28 indicating that the majority of stock price
performance values lie close to this mean. While the raw stock price performance
variable exhibits moderately negative skew, all three variables are relatively normally
distributed. The mean stock beta value of 0.97 indicates an average, normal level of
risk amongst the 25 companies.
Table 6.18 – Descriptive Statistics for Stock Price Performance: Overproviders
Extreme Abnormal Overproviders Mean Std
Dev Min Max Skew
Raw Stock Price Performance (%)* 3.89 0.24 -34.5 67.10 0.63
Stock Beta: Risk Variable 1.09 0.65 0.13 2.81 0.95
Natural Log of Total Assets: Size Variable 3.13 0.55 2.18 3.99 -0.12
All Percentages (Indicated*) are converted to Percentage Point Scores through Multiplication – (X 100)
As indicated in Table 6.18 above, the 25 most extreme abnormal overproviders
experienced an average post financial year end stock price performance of +3.9%.
Chapter 6 Research Findings
___________________________________________________________________________
66
The standard deviation value of 0.24 suggests that the majority of stock price
performance values lie close to this mean, while the mean stock beta value of 1.09
indicates an average, slightly greater than normal level of risk amongst the 25
companies. Notably, there is a significant (15%) difference between the average stock
price performance of the 25 most extreme abnormal underproviders and
overproviders, as indicated in Figure 6.7 below.
Figure 6.7 – Mean Stock Price Performance Post Financial Year End
Evidently, the 25 FTSE 350 companies with extreme abnormal underprovision for
credit loss on trade receivables experienced an inferior stock price performance, post
financial year end, relative to the 25 FTSE 350 companies with extreme abnormal
overprovision for credit loss on trade receivables. H16 is therefore accepted.
6.5.2 Multiple Regression Analysis: Stock Price Performance Significance
OLS multiple regression analysis was conducted to examine the extent of multivariate
relationships between the dependent variable: Stock Price Performance
(Methodology: Section 5.9.2) and extreme abnormal provision for credit less on trade
receivables. Both the Beta Value and Natural Logarithm of Total Assets variables
were included to control for both risk and size. The corresponding results are
contained in Appendix D.
Chapter 6 Research Findings
___________________________________________________________________________
67
When considered in aggregate, the results of regression one through regression three
inclusive (Appendix D) indicate that extreme abnormal underprovision for credit loss
on trade receivables significantly explains resultant post financial year end stock price
performance, while extreme abnormal overprovision is not significantly explanatory
in this regard.
Indeed, the extent of extreme abnormal underprovision for credit loss on trade
receivables exhibits a significant positive (5% level) relationship with resultant stock
price performance, where a 1% increase in earnings quality (reduction in the extent of
abnormal underprovision) results in a 0.416% more positive stock price performance.
6.6 Conclusion
The results of univariate and multivariate analyses, as presented in this chapter,
confirm the widespread existence of mean abnormal underprovision for credit loss on
trade receivables amongst FTSE 350 companies during the latest period.
A statistically significant negative association between the Change in Gearing and
Change in Gross Margin variables and the direction of abnormal provision for credit
loss on trade receivables is identified, with a significant positive association in the
case of the Proportion of INEDs to the Total Board variable. The results also confirm
that FTSE 350 companies with extreme abnormal underprovision for credit loss on
trade receivables experienced an inferior mean post financial year end stock price
performance, relative to those FTSE 350 companies with extreme abnormal
overprovision.
Chapter 7
Discussion
“One can’t say that figures lie. But figures, as used in financial
arguments, seem to have the bad habit of expressing a small part of the
truth forcibly”.
Fred Schwed Jr. (1940)
Chapter 7 Discussion
___________________________________________________________________________
69
CHAPTER SEVEN
DISCUSSION
7.1 Introduction
This chapter examines the implications of the research findings relative to regulation,
the previously discussed empirical research and accounting and economic theory
underlying earnings quality and earnings management. The major themes arising from
the findings are discussed sequentially, in the context of the research questions and
research objectives.
Chapter 7 Discussion
___________________________________________________________________________
70
7.2 Magnitude of Abnormal Provision for Credit Loss on Trade Receivables
7.2.1 Provisioning Activity at Variance with Credit Risk Environment
The aim of research objective one was to quantify the existence, direction and
magnitude of abnormal provision for credit loss on trade receivables amongst
FTSE 350 companies. Amidst a continuously challenging economic environment,
with significant pressure on corporate earnings (Bloomberg, 2012), the mean relative
increase in total gross trade receivables of 12.0 per cent suggests that companies are
increasing the extent of their credit sales, in itself a measure that increases the
inherent risk of default or credit loss, most likely to maintain or stimulate growth in
top line revenues. The mean increase of only 0.88 days in the average trade
receivables collection period also reinforces the concept of elevated credit risk, where
any extension in trade credit has been heavily restricted. Moreover, the mean relative
increase of only 2.1 per cent in the provision for credit loss on trade receivables
provides strong evidence of mean abnormal underprovision of -9.9 per cent, with 138
of the 204 FTSE 350 companies underproviding, despite an environment where
European corporates expect an even stronger deterioration in credit risk during 2012
(Atradius, 2012).
7.2.2 Increased Credit Delinquency: Downside Risk of Elevated Write-Offs
During the latest period, the mean level of trade receivables past due, but not impaired
(where disclosed) amounted to £ 138.8 million, compared with a mean of £ 128.0
million in the previous period, representing a relative increase of 8.4 per cent.
Moreover, the average rate of provision for credit loss on trade receivables amongst
the 204 FTSE 350 companies declined to (5.98%) from (6.41%) in the previous
period. Despite the categorisation of non-impairment, there has been a clear increase
in the level of credit delinquency amongst the 204 FTSE 350 companies, rendering
the identified underprovision all the more suspect. With such widespread relative
underprovision, the downside risk arising from instances of elevated credit loss in the
near term is significant. Should credit risk deteriorate further as anticipated, there is
an underlying risk of increased write-offs and a negative impact upon corporate
earnings as a result of increased provisioning activity - given the current level of
abnormal underprovision.
Chapter 7 Discussion
___________________________________________________________________________
71
7.2.3 Regulatory Considerations
While no consistent approach to credit risk assessment was noted amongst the 204
FTSE 350 companies, significant divergence from the disclosure requirements of
IFRS 7 was noted throughout, particularly amongst those companies in the mining
and natural resources sectors. Of the final sample of 204 FTSE 350 companies, only
98 provided detail regarding trade receivables past due and specifically impaired in a
manner consistent with IFRS 7, while 178 provided detail regarding trade receivables
past due and not impaired in a manner consistent with IFRS 7. Such non-compliance
and the resultant information asymmetry provides ample opportunity for manipulation
of the provision for credit loss on trade receivables in the context of earnings
management.
Extensive regulation of the provision for credit loss on trade receivables by the IASB
is clearly impractical, as the impairment of trade receivables is inherently subjective
and unique to each company or sector type. However, enhanced disclosure
compliance with IFRS 7 is fundamentally important, given the extensive
abnormalities identified in this study, where some 90.1 per cent of the variation in the
change in provision for credit loss on trade receivables is explained by factors beyond
the relative change in gross trade receivables.
7.3 Determinants of Abnormal Provision for Credit Loss on Trade Receivables
The aim of research objective two was to develop a multivariate OLS regression
model that examines the applicability of previously identified and alternative
determinants of earnings management to abnormal provision for credit loss on trade
receivables amongst FTSE 350 companies.
7.3.1 Capital Market Variables – Limited Evidence
The significance of capital market determinants, including analyst earnings
expectations, is well documented in previous earnings management research (Healy
and Wahlen, 1998). However, the findings of this study provide only limited evidence
of a relationship between these variables and the direction of abnormal provision for
credit loss on trade receivables.
Chapter 7 Discussion
___________________________________________________________________________
72
Both the Consensus EPS Growth % and Earnings (EPS) Surprise % variables exhibit
an insignificant relationship with the direction of abnormal provision for credit loss on
trade receivables, which is surprising, given their significance in previous studies
including Payne and Robb (1997) and Kasznik (1999). Although the negative
association between the Consensus EPS Growth % variable and the direction of
abnormal provision for credit loss on trade receivables is insignificant, this result
provides limited support for the conception in agency theory that agents engage in
earnings management in order to comply with consensus forecasts (Palliam and
Shalhoub, 2003), where the extent of earnings quality decreases (abnormal
underprovision increases) as consensus earnings forecasts increase.
The insignificant, yet positive association between the Earnings (EPS) Surprise %
variable and the direction of abnormal provision, while surprising, provides limited
evidence that instances of increasing earnings surprise are complemented with
instances of abnormal overprovision for credit loss on trade receivables. While the
204 FTSE 350 companies are clearly not utilising abnormal underprovision to support
instances of positive earnings surprise, the mean earnings surprise value of (0.33%)
indicates that, on average, the total sample exceeded analyst earnings expectations,
consistent with Lopez and Rees (2002).
7.3.2 Contractual Variables – Significant Evidence
Contractual determinants of earnings management examined in this study comprise
the Existence of a Bonus Plan, Executive Incentive Remuneration and the Change in
Gearing. In both univariate and multivariate analyses, the Change in Gearing exhibits
a significant negative relationship with the direction of abnormal provision for credit
loss on trade receivables, at the 5% and 10% levels respectively.
This result, consistent with that hypothesised, indicates that increasing levels of
gearing significantly explain abnormal underprovision for credit loss on trade
receivables, providing support for the earnings inflation motive attached to abnormal
underprovision, where earnings inflation procedures are adopted to reduce the risk of
violating debt covenants (Dechow et al, 1996).
Chapter 7 Discussion
___________________________________________________________________________
73
While demonstrating consistency with DeFond and Jimbalvo (1994) and Sweeney
(1994), this result provides strong support for the debt hypothesis of positive
accounting theory, where: there is a positive relationship between an increasing debt
to equity ratio (level of gearing) and the likelihood of the adoption of earnings
inflation procedures specifically (Watts and Zimmerman, 1986). Amongst the 204
FTSE 350 firms, this result indicates that abnormal underprovision for credit loss on
trade receivables is utilised as an instrument for earnings inflation where the level of
gearing increases. However, it may also be that such abnormal underprovision is
utilised to enhance debt holders’ perceptions of the company’s prospects or to engage
in impression management more broadly.
Notably, neither the Existence of a Bonus Plan nor Executive Incentive Remuneration
variables exhibit a significant relationship with the direction of abnormal provision
for credit loss on trade receivables, contrasting sharply with prior research including
Healy and Wahlen (1998), Chen (2006) and the management compensation
hypothesis of positive accounting theory (Watts and Zimmerman, 1986). However,
given that 99 per cent of the 204 FTSE 350 companies operated bonus or incentive
remuneration related plans during the latest period; this variable is clearly no longer
robust in earnings management research. Nevertheless, the insignificant, yet negative
association between the Executive Incentive Remuneration variable and the direction
of abnormal provision for credit loss on trade receivables provides limited evidence
supporting Balsam (1998) and Gaver and Gaver (1998) – where the extent of earnings
quality decreases (abnormal underprovision increases) as executive incentive
remuneration increases.
7.3.3 Performance Variables – Significant Evidence
The Change in Gross Margin variable is found to have a significant negative
relationship with the direction of abnormal provision for credit loss on trade
receivables in both univariate and multivariate analyses at the 5% level, indicating
that instances of an increasing gross margin significantly explain resultant abnormal
underprovision for credit loss on trade receivables.
Chapter 7 Discussion
___________________________________________________________________________
74
While consistent with the findings of Zhang (2006), this result provides additional
support for the earnings inflation motive adopted throughout this study, clearly
highlighting a motivation to maintain positive top line performance through to the
final earnings performance, utilising abnormal underprovision in this regard.
However, neither the Change in Net Margin nor Change in Trade Receivables Days
variables exhibit a statistically significant relationship with the direction of abnormal
provision for credit loss on trade receivables.
7.3.4 Governance Variables – Significant Evidence
The intervening and mitigation effects of robust corporate governance structures on
earnings management are widely acknowledged (Sebahattin and Harlan, 2009).
Governance specific determinants of earnings management examined in this study
comprise the Proportion of INEDs to the Total Board, Proportion of INEDs to the
Audit Committee, No. of Audit Committee Meetings and No. of Governance Non
Compliance Issues. Consistent with Lin et al (2006), no significant relationship is
found between the No. of Audit Committee Meetings variable and the direction of
abnormal provision for credit loss on trade receivables. However, the direction of the
relationship between all four variables and abnormal provision for credit loss on trade
receivables is consistent with that hypothesised, while the Proportion of INEDs to the
Total Board variable exhibits a significant positive (5% level) relationship with the
direction of abnormal provision for credit loss on trade receivables.
These findings provide strong evidence of decreasing earnings inflation activity as the
robustness of corporate governance structures increase, consistent with Dechow et al
(1996), Beasley (1996), Peasnell et al (2005) and Sebahattin and Harlan (2009).
Indeed, with resultant abnormal overprovision for credit loss on trade receivables as
the proportion of INEDs to the total board of directors increases (categorised as
prudence rather than earnings deflation in an environment of elevated credit risk), this
study provides a strong impetus for continued convergence towards corporate
governance best practice, where the board of directors comprises of majority INEDs.
Chapter 7 Discussion
___________________________________________________________________________
75
A strong trend towards majority INED board composition has developed in recent
years (Grant Thornton, 2011) and this study confirms that the average board of
directors of the 204 FTSE 350 companies comprised of 53.9% INEDs during the
latest period. Nevertheless, amongst the 204 FTSE 350 firms, 52 reported board
structures that comprised of less than 50% INEDs, while eight reported audit
committee structures that comprised of less than three INEDs. The clear support for
the mitigating effects of robust corporate governance structures on earnings inflation
activity (via abnormal underprovision for credit loss on trade receivables specifically)
highlights the importance of such mechanisms in protecting stakeholders from
earnings management activity.
7.3.5 Auditor Variables – Mixed Evidence
Large auditors with industry specialist knowledge are widely considered to be more
effective in constraining the earnings management activity of their clients (Francis
and Krishnan, 1999). The findings of this study, upon initial examination, provide
some support in this regard, with an apparent significant positive (5% level)
relationship between the Auditor Type variable and the direction of abnormal
provision for credit loss on trade receivables. However, as detailed in Chapter 6, the
abnormal distribution of the Auditor Type (categorical) variable has skewed the
results at the multivariate level and this result is therefore disregarded. Given that 198
or 97 per cent of the final sample were audited by Big 4 firms, this variable clearly
lacks robustness in earnings management research. Contrasting with Frankel et al
(2002), no significant relationship between the Audit Specific Fee(s) variable and the
direction of abnormal provision for credit loss on trade receivables is exhibited.
7.3.6 Non Hypothesised Factors
The Adj. R Square values resulting from regression analyses examining research
objective two range from 1.6 – 8.9 per cent, with a significant extent of the variation
in the direction of abnormal provision for credit loss on trade receivables therefore
explained by factors omitted from the model employed. This result is not surprising
however, given that Dechow et al (2011) document the lack of power in earnings
management models and the relatively low Adj. R Square values recorded in prior
earnings management research; including Frankel et al (2002) and Chen (2006).
Chapter 7 Discussion
___________________________________________________________________________
76
The overall decision to manage earnings (abnormally provide for credit loss on trade
receivables) is a complex one; affected by the broader macroeconomic environment
and sector specific variables but also significantly by the economic and political
circumstances of a firm (Peltier-Rivest and Swirsky, 2000). While both the positive
accounting and agency theories provide a firm foundation for the determinants of
earnings management, the decision to manage earnings ultimately comprises many
variables that are simply non-linear and not quantifiable but qualitative in nature.
7.4 Stock Price Performance of Extreme Abnormal Providers
7.4.1 Evidence Supporting the Efficient Market Hypothesis
The aim of research objective three was to examine the individual and aggregate stock
price performance of the most extreme abnormal providers for credit loss on trade
receivables (both under and over providers) over a specified post financial year end
period. The results provide strong evidence that, of the 204 FTSE 350 companies,
those with the most extreme abnormal underprovision for credit loss on trade
receivables experienced an average negative (-11.1%) post financial year end raw
stock price performance, significantly inferior to the performance of the most extreme
abnormal overproviders (+3.9%). This result is directly consistent with Holthausen et
al (1995), Chan et al (2001) and Dechow et al (2007) – where companies with low
quality earnings experience a subsequent negative or inferior stock price performance
post issuance of their financial statements. Contrasting with Sloan (1996), this result
provides support for the efficient market hypothesis, indicating that capital market
participants identify instances of low quality earnings and discount stocks
accordingly, or discount stocks where abnormalities have already been identified.
7.4.2 Capital Markets: A Potential Instrument for Effective Regulation
This result further suggests that capital markets may represent a potential instrument
for regulating accounting abnormalities, somewhat mitigating the effects of
accounting standard non-compliance and earnings management, with the discounting
of stocks where abnormal underprovision is exhibited acting as a warning signal to
prospective investors. Moreover, the abnormal underprovision identified in this study
may also be complemented by a series of additional earnings inflation accounting
abnormalities that further explain the subsequent inferior stock price performance.
Chapter 7 Discussion
___________________________________________________________________________
77
7.4.3 Non Hypothesised Factors
Notably, there is a slight difference between the average risk and size profiles of those
extreme abnormal underproviders and overproviders, with extreme abnormal
underproviders exhibiting slightly lower average risk and smaller size profiles relative
to those extreme abnormal overproviders. While neither of these factors was
hypothesised as a determinant of abnormal provision, these results suggest that in
instances of extreme abnormal underprovision; smaller companies with lower risk
profiles, arguably subject to reduced capital market scrutiny and monitoring, may be
availing of such reduced supervision to actively engage in earnings management.
7.5 Conclusion
This chapter, discussing the major themes and findings of this study, has identified a
significant link between positive accounting theory, prior empirical evidence and
extant earnings management practice. Theoretical foundations conceptualised some
forty years ago, including the debt hypothesis of positive accounting theory remain
robust in explaining the motives for earnings management activity. While presenting
strong evidence of widespread mean abnormal underprovision for credit loss on trade
receivables, this chapter also highlights the potential regulatory ability of capital
markets, possibly mitigating the effects and costs of earnings management activity.
Chapter 8
Conclusion
“Some regulation will be necessary, some changes in accounting rules”.
Henry Paulson (2008) – U.S. Department of the Treasury
Chapter 8 Conclusion
___________________________________________________________________________
79
CHAPTER EIGHT
CONCLUSION
8.1 Introduction
This final chapter draws the dissertation to a close, summarising the major findings
and conclusions reached. While making recommendations for practitioners including
auditors, accounting standard setters and capital market participants, this chapter also
identifies opportunities for further research.
Chapter 8 Conclusion
___________________________________________________________________________
80
8.2 Research Questions, Research Objectives and Findings
The following research questions have provided the basis for the analysis undertaken
throughout this dissertation:
RQ 1 – What is the magnitude and what are the determinants of abnormal provision
for credit loss on trade receivables amongst FTSE 350 companies?
RQ 2 – What is the capital (stock) market response to instances of extreme abnormal
provision for credit loss on trade receivables amongst FTSE 350 companies?
The three primary research objectives analysed and resultant major findings include:
Objective One: Key Findings
Objective: to quantify the existence, direction and magnitude of abnormal provision
for credit loss on trade receivables amongst FTSE 350 companies. The mean level of
abnormal underprovision of -9.9 per cent provides strong evidence of provisioning
practice at sharp variance with the current environment of elevated credit risk. The
results also indicate that some 90.1 per cent of the variation in the change in provision
for credit loss on trade receivables is explained by factors other than the relative
change in total gross trade receivables.
Objective Two: Key Findings
Objective: to develop a multivariate OLS regression model that examines the
applicability of previously identified and alternative determinants of earnings
management to abnormal provision for credit loss on trade receivables amongst FTSE
350 companies.
The results indicate that increasing gross margins significantly explain abnormal
underprovision. Strong support is also exhibited for the debt hypothesis of positive
accounting theory as a determinant of earnings management, where earnings inflation
through abnormal underprovision is undertaken to avoid the risk of violating debt
covenants. Significant support for the intervening and mitigating effects of robust
corporate governance structures on earnings inflation via abnormal underprovision is
also identified, where the board of directors is comprised of an increasing proportion
of INEDs.
Chapter 8 Conclusion
___________________________________________________________________________
81
Objective Three: Key Findings
Objective: to examine the individual and aggregate stock price performance of the
most extreme abnormal providers for credit loss on trade receivables over a specified
post financial year end period. The results provide support for the efficient market
hypothesis that stock prices fully reflect all publicly available information, where
those FTSE 350 firms with extreme abnormal underprovision experience an inferior
post financial year end stock price performance. Moreover, in regulating accounting
abnormalities through the discounting of stock prices, capital markets are identified as
a potential instrument for mitigating the effects of accounting standard
non-compliance and earnings management.
8.3 Limitations of Research
The findings and conclusions reached throughout this study are subject to a number of
important limitations. It is possible that the attachment of prudence rather than an
earnings deflation motive to abnormal overprovision disregards a considerable degree
of earnings management activity. Although testing procedures validate compliance
with the underlying assumptions of OLS regression analysis, the justification for the
use of the ordinary least squares estimation is diminished where any additional
conditions are not observed entirely. This study considers the most recent change in
the provision for credit loss on trade receivables amongst a refined sample of 204
FTSE 350 companies over the latest available financial period. It does not consider
additional variables, including discretionary accruals, through which earnings
management may be exercised, nor does it adopt a longitudinal approach.
8.4 Recommendations for Practitioners
8.4.1 Auditors
This study highlights the potential for earnings inflation via the provision for credit
loss on trade receivables specifically. In circumstances where audit clients exhibit
increased levels of gearing, increasing gross margins or deficient corporate
governance structures, auditors should dedicate additional resources to examining
both anticipated credit losses on trade receivables and broader provisioning activities.
While not significantly material in isolation, multiple instances of undetected
manipulation via abnormal provisioning can serve to distort overall earnings.
Chapter 8 Conclusion
___________________________________________________________________________
82
8.4.2 International Accounting Standards Board – Standard Setters
The significant level of non-compliance with the disclosure requirements of IFRS 7
recorded in this study and the resultant information asymmetry provides ample
opportunity for manipulation. Amidst the 2008 financial crisis, Henry Paulson, former
U.S. Secretary of the Treasury, stated that: “Some regulation will be necessary, some
changes in accounting rules”. In a similar vein, while complete regulation of the
provision for credit loss on trade receivables is impractical, the abnormalities
identified in this study provide a strong impetus for at least limited regulatory change.
A mandatory, standardised IFRS 7 reporting format or template, drafted by the IASB,
could serve well in eliminating the current non-compliance, while simultaneously
reducing the potential for manipulation.
8.4.3 Capital Market Participants
Despite strong support for the efficient market hypothesis in this study, both
shareholders and corporate stakeholders should apply detailed scrutiny to the
governance structures and board level composition of companies. Such scrutiny can
help to prevent against adverse asset allocation, investment and corporate lending
decisions – where these decisions are based upon higher quality earnings that are
supported by the existence of robust governance structures.
8.5 Recommendations for Future Research
This study may be expanded in a number of ways. Firstly, a longitudinal analysis of
the provision for credit loss on trade receivables could be conducted, to determine
whether multi-period abnormalities are exhibited and the extent to which the phase of
the business cycle impacts the extent of these abnormalities.
Those companies exhibiting extreme abnormal underprovision for credit loss on trade
receivables could be examined across a broader spectrum of earnings quality
indicators to ascertain the extent of overall earnings quality and earnings management
activity.
Finally, both financial and stock price performance in the years subsequent to
abnormal underprovision could be examined, to determine whether it serves as a
precursor to long run declining performance or financial distress.
Chapter 8 Conclusion
___________________________________________________________________________
83
8.6 Conclusion
Writing in 1940 about the practices of Wall Street investors, Fred Schwed Jr. stated
that: “One can’t say that figures lie. But figures, as used in financial arguments, seem
to have the bad habit of expressing a small part of the truth forcibly”. Over 70 years
later, this statement holds significant truth given the abnormalities identified in this
study. In a post 2008 financial crisis world, with corporate financial information more
readily available than ever before and accounting regulation once again in the
spotlight, the opportunity to engage in earnings management is ever diminishing.
However, this study confirms the existence and potential for earnings management
through abnormal provision for credit loss on trade receivables. While excessive
regulation is impractical and may result in disproportionate monitoring costs, this
study provides a strong impetus for at least limited regulatory change.
Regulation, the macroeconomic environment and firm specific variables all fluctuate
over time, however; the fundamentals underlying both the agency and positive
accounting theories remain constant. Where conditions are compliant with either of
these theories, there exists a high risk of earnings management activity through
abnormal provision for credit loss on trade receivables; with potential adverse
implications across a broad spectrum of corporate stakeholders.
___________________________________________________________________________
___________________________________________________________________________
References
References
___________________________________________________________________________
85
REFERENCES
Anderson, K. L. and Yohn, T. L. (2002) ‘The Effect of 10-K Restatements on Firm
Value, Asymmetries, and Investors' Reliance on Earnings’ [Online]. Available at:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=332380 (Accessed 12 January
2012).
Atradius (2012) ‘Atradius Payment Practices Barometer - Core Results of Western
Europe’ [Online]. Available at: http://global.atradius.com/images/stories/Publications/
payment_pratices/Round11/PPB11_HQ_en_fin.pdf (Accessed 15 May 2012).
Ayres, F. L. (1994) 'Perceptions of earnings quality: what managers need to know',
Strategic Finance, Vol. 75, No. 2, pp. 27-29.
Balsam, S. (1998) 'Discretionary accounting choices and CEO compensation',
Contemporary Accounting Research, Vol. 15, No. 3, pp. 229-252.
Beasley, M. S. (1996) 'An Empirical Analysis of the Relation Between the Board of
Director Composition and Financial Statement Fraud', The Accounting Review,
Vol. 71, No. 4, pp. 443-465.
References
___________________________________________________________________________
86
Bellovary, J. L., Giacomino, D. E. and Akers, M. D. (2005) 'Earnings Quality: It's
Time to Measure and Report', The CPA Journal, Vol. 75, No. 11, pp. 32-37.
Bitner, L. (2005) 'Earnings Management: How to Identify Contributing Factors',
Commercial Lending Review, Vol. 20, No. 5, pp. 25-28.
Bloomberg (2012) 'Euro Crisis Hits Profits Globally As P&G Cuts Forecast' [Online].
Available at: http://www.bloomberg.com/news/2012-06-24/euro-crisis-hits-profits-
globally-as-p-g-cuts-forecast.html (Accessed 07 July 2012).
Bricker, R., Previts, G., Robinson, T. and Young, S. (1995) 'Financial analyst
assessment of company earnings quality', Journal of Accounting, Auditing and
Finance, Vol. 10, pp. 541-554.
Burgstahler, D. and Dichev, I. (1997) 'Earnings Management to Avoid Earnings
Decreases and Losses', Journal of Accounting and Economics, Vol. 24, No. 1,
pp. 99-126.
Caylor, M. L. (2009) ‘Strategic Revenue Recognition to Achieve Earnings
Benchmarks’ [Online]. Available at: http://papers.ssrn.com/sol3/papers.
cfm?abstract_id =885368 (Accessed 15 April 2012).
Chan, K., Jegadeesh, N., Louis, C. and Lakonishok, J. (2001) ‘Earnings Quality and
Stock Returns: The Evidence from Accruals’ [Online]. Available at:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=259691 (Accessed 15 January
2012)
Chen, C. (2006) ‘Earnings Management, Earnings Manipulation: Evidence from
Taiwanese Corporations’ [Online]. Available at: http://aaahq.org/audit/midyear/
07midyear/papers/Chen%20_EarningsManagementEarningsManipulation.pdf
(Accessed 04 December 2011).
Collins, J. H., Shackleford, D. A. and Wahlen, J. M. (1995) 'Bank differences in the
coordination of regulatory capital, earnings and taxes', Journal of Accounting
Research, Vol. 33, No. 2, pp. 263-291.
References
___________________________________________________________________________
87
Colson, R., Bloomfield, R., Christensen, E., Jamal, K. and Moehrle, S. (2010)
'Response to the Financial Accounting Standards Board's and the International
Accounting Standards Board’s Joint Discussion Paper Entitled: Preliminary Views on
Revenue Recognition In Contracts With Customers', Accounting Horizons, Vol. 24,
No. 4, pp. 689-702.
Curcio, D. and Hasan, I. (2008) ‘Earnings - and Capital Management and Signalling:
The Use of Loan-Loss Provisions by European Banks’, Paper presented at the 2009
Annual European Financial Management Association Conference, 24-27 June,
Universita Bocconi, Milan, Italy.
DeAngelo, H., DeAngelo, L. and Skinner, D. (1994) 'Accounting Choice in Troubled
Companies', Journal of Accounting and Economics, Vo. 17, No. 1, pp. 113-143.
DeAngelo, L. E. (1988) 'Managerial Competition, Information Costs and Corporate
Governance: The use of accounting performance measures in proxy contests', Journal
of Accounting and Economics, Vol. 10, No. 1, pp. 3-36.
DeAngelo, L. E. (1986) 'Accounting Numbers as Market Valuation Substitutes: A
Study of Management Buyouts of Public Stockholders', The Accounting Review,
Vol. 61, No. 3, pp. 400-420.
Dechow, P. M. and Dichev, I. D. (2002) 'The quality of accruals and earnings: the role
of estimation errors', The Accounting Review, Vol. 77, (Supplement), pp. 35-59.
Dechow, P. M., Ge, W., Larson, C. R. and Sloan, R. G. (2007) ‘Predicting Material
Accounting Manipulations’ [Online]. Available at: http://papers.ssrn.com/sol3/
papers.cfm?abstract_id=997483 (Accessed 01 June 2012).
Dechow, P. M., Hutton, A. P., Kim, J. H. and Sloan, R. G. (2011) 'Detecting Earnings
Management: A New Approach' [Online]. Available at: http://papers.ssrn.com/sol3/
papers.cfm?abstract_id=1735168 (Accessed 07 January 2012).
References
___________________________________________________________________________
88
Dechow, P. M. and Skinner, D. J. (2000) 'Earnings Management: Reconciling the
View of Accounting Academics, Practitioners, and Regulators', Accounting Horizons,
Vol. 14, No. 2, pp. 235-250.
Dechow, P. M. and Sloan, R. G. (1991) 'Executive Incentives and the Horizon
Problem: An Empirical Investigation', Journal of Accounting and Economics,
Vol. 14, No. 1, pp. 51-89.
Dechow, P. M., Sloan, R. G. and Sweeney, A. P. (1995) 'Detecting Earnings
Management', The Accounting Review, Vol. 70, No. 2, pp. 193-225.
Dechow, P. M., Sloan, R. G. and Sweeney, A. P. (1996) 'Causes and Consequences of
Earnings Manipulation: An Analysis of Firms Subject to Enforcement Actions by the
SEC', Journal of Contemporary Accounting Research, Vol. 13, No. 1, pp. 1-36.
DeFond, M. L. and Jimbalvo, J. (1994) 'Debt covenant violation and manipulation of
accruals', Journal of Accounting and Economics, Vol. 17, No. 2, pp. 145-176.
Degeorge, F., Patel, J. and Zeckhauser, R. (1999) 'Earnings Management To Exceed
Thresholds', The Journal of Business, Vol. 72, No. 1, pp. 1-33.
Duncan, J. R. (2002) 'Quality of earnings and accounting estimates: a case of
decisions', Issues in Accounting Education, Vol. 17, No. 4, pp. 401-409.
Dutta, S. and Zhang, X. J. (2000) ‘Revenue Recognition in a Multi-Period Agency
Setting’, Journal of Accounting Research, Vol. 40, No. 1, pp. 67-83.
EFMA (2012) ‘European Credit Risk Outlook: Results of the Fourth European Credit
Risk Managers Survey’ [Online]. Available at: http://www.efma.com/
ressources/studies/2012/1-OKP4X_E_study.pdf (Accessed 01 June 2012).
Erickson, M. and Wang, S. (1999) 'Earnings Management by Acquiring Firms in
Stock for Stock Mergers', Journal of Accounting and Economics, Vol. 27, No. 2,
pp. 149-176.
References
___________________________________________________________________________
89
London Stock Exchange (2012) 'FTSE 350' [Online]. Available at:
http://www.londonstockexchange.com/exchange/prices-and-markets/stocks/indices/
summary/summary-indices-constituents.html?index=NMX (Accessed 01 June 2012).
Francis, J. R. and Krishnan, J. (1999) 'Accounting Accruals and Auditor Reporting
Conservatism', Contemporary Accounting Research, Vol. 16, No. 1, pp. 135-165.
Frankel, R. M., Johnson, M. F. and Nelson, K. K. (2002) 'The Relation between
Auditors' Fees for Nonaudit Services and Earnings Management', The Accounting
Review, Vol. 77, (Supplement), pp. 71-105.
Gaver, J. J. and Gaver, K. M. (1998) 'The relation between nonrecurring accounting
transactions and CEO cash compensation', Accounting Review, Vol. 73, No. 2,
pp. 235-253.
Grant Thornton (2011) 'Corporate Governance Review 2011' [Online]. Available at:
http://www.grant-thornton.co.uk/pdf/Corporate_Governance_Review_2011.pdf
(Accessed 05 June 2012).
Greenspan, A. (2007) The Age of Turbulence: Adventures in a New World, United
States of America: Penguin Press.
Habib, A. and Hansen, J. (2008) 'Target Shooting: Review of Earnings Management
around Earnings Benchmarks', Journal of Accounting Literature, Vol. 27, pp. 25-70.
Healy, P. M. (1985) 'The Effect of Bonus Schemes on Accounting Decisions', Journal
of Accounting and Economics, Vol. 7, No. 1-3, pp. 85-107.
Healy, P. M. and Wahlen, J. M. (1998) ‘A Review Of The Earnings Management
Literature And Its Implications For Standard Setting’, Accounting Horizons, Vol. 13,
No. 4, pp. 365-383.
Holthausen, R., Larcker, D. and Sloan, R. (1995) 'Annual bonus schemes and the
manipulation of earnings', Journal of Accounting and Economics, Vol. 19, No. 1,
pp. 29-74.
References
___________________________________________________________________________
90
Howell, D. C. (2009) Statistical Methods for Psychology, Belworth, CA: Wadsworth
Publishing.
IASB (2011) IFRS - Part A: The Conceptual Framework and Requirements, London,
United Kingdom: International Accounting Standards Board Publications.
Jeffrey, L., Callen, S., Robb, S. and Segal, D. (2008) 'Revenue Manipulation and
Restatements by Loss Firms', Auditing: A Journal of Practice and Theory, Vol. 27,
No. 2, pp. 1-29.
Jensen, M. C. and Meckling, W. H. (1976) 'Theory of the Firm: Managerial
Behaviour, Agency Costs and Ownership Structure', Journal of Financial Economics,
Vol. 3, No. 4, pp. 305-360.
Jones, J. (1991) 'Earnings Management During Import Relief Investigations', Journal
of Accounting Research, Vol. 29, No. 2, pp. 193-228.
Jordan, C. E., Clark, S. J. and Hames, C. C. (2010) 'The Impact Of Audit Quality On
Earnings Management To Achieve User Reference Points in EPS', Journal of Applied
Business Research, Vol. 26, No. 1, pp. 19-30.
Jouber, H. and Fakhfakh, H. (2011) 'Earnings management and board oversight: an
international comparison', Managerial Auditing Journal, Vol. 27, No. 1, pp. 66-86.
Kasznik, R. (1999) 'On the Association between Voluntary Disclosure and Earnings
Management', Journal of Accounting Research, Vol. 37, No. 1, pp. 57-81.
Krishnan, G. V. (2003) 'Does Big 6 Auditor Industry Experience Constrain Earnings
Management?', Accounting Horizons, Vol. 17, (Supplement), pp. 1-16.
Lev, B. and Thiagarajan, B. (1993) 'Fundamental Information Analysis', Journal of
Accounting Research, Vol. 31, No. 2, pp. 190-215.
Lin, J. W., Li, J. F. and Yang, J. S. (2006) 'The effect of audit committee performance
on earnings quality', Managerial Auditing Journal, Vol. 21, No. 9, pp. 921-933.
References
___________________________________________________________________________
91
Lopez, T. and Rees, L. (2002) 'The effect of beating and missing analysts' forecasts on
the information content of unexpected earnings', Journal of Accounting, Auditing and
Finance, Vol. 17, No. 2, pp. 155-184.
Macve, R. and Serafeim, G. (2006) ‘Deprival Value VS Fair Value Measurement For
Contract Liabilities in Resolving the Revenue Recognition Conundrum: Towards a
General Solution’ [Online]. Available at: http://www2.lse.ac.uk/accounting/faculty
AndStaff/liabrevrec300606ssrn.pdf (Accessed 29 November, 2011).
McNichols, M. F. (2000) 'Research Design Issues in Earnings Management Studies',
Journal of Accounting and Public Policy, Vol. 19, No. 4-5, pp. 313-345.
McNichols, M. F. and Wilson, P. G. (1988) 'Evidence of Earnings Management from
the Provision for Bad Debts', Journal of Accounting Research, Vol. 26, pp. 1-31.
Mohammady, A. (2010) 'Earnings Quality Constructs and Measures', The Business
Review, Vol. 16, No. 1, pp. 277-283.
Morris, R. D. (1987) 'Signalling, Agency Theory and Accounting Policy Choice',
Accounting and Business Research, Vol. 18, No. 69, pp. 47-56.
Newbold, P. (1988) Statistics for Business and Economics, 2nd Edn. United States of
America: Prentice-Hall International.
O' Mahony, O. (2010) An Analysis of the Determinants of Voluntary Impairment
Disclosure in the Annual Reports of UK and Irish Companies, MBS Dissertation,
Waterford Institute of Technology.
Palliam, R. and Shalhoub, Z. (2003) 'The Phenomenology of Earnings Management
within the Confines of Agency Theory', International Journal of Value - Based
Management, Vol. 16, No. 1, pp. 65-88.
Payne, J. and Robb, S. (1997) ‘Earnings Management: The effect of ex ante earnings
expectations’ [Online]. Available at: http://papers.ssrn.com/sol3/papers.cfm?
abstract_id=2859 (Accessed 18 January 2012).
References
___________________________________________________________________________
92
Peasnell, K. V., Pope, P. F. and Young, S. (2005) 'Board Monitoring and Earnings
Management: Do Outside Directors Influence Abnormal Accruals?', Journal of
Business Finance and Accounting, Vol. 32, No. 7-8, pp. 1311-1346.
Peltier-Rivest, D. (1999) 'The Determinants of Accounting Choices in Troubled
Companies', Quarterly Journal of Finance and Accounting, Vol. 38, No. 4, pp. 28-44.
Peltier-Rivest, D. and Swirsky, S. (2000) 'Earnings Management in Healthy Firms',
Quarterly Journal of Finance and Accounting, Vol. 39, No. 4, pp. 21-37.
Perry, S. E. and Williams, T. H. (1994) 'Earnings management preceding management
buyout offers', Journal of Accounting and Economics, Vol. 18, No. 2, pp. 157-179.
Ricci, C. W. (2011) 'Manipulating Receivables: A Comparison Using the SEC's
Accounting and Auditing Enforcement Releases', Journal of Applied Business and
Economics, Vol. 12, No. 5, pp. 35-44.
Schipper, K., Franics, J., LaFond, R. and Olsson, P. (2003) ‘Earnings Quality and the
Pricing Effects of Earnings Patterns’ [Online]. Available at:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=414142 (Accessed 15 January
2012).
Schipper, K. and Vincent, L. (2003) 'Earnings Quality', Accounting Horizons, Vol. 17,
(Supplement), pp. 97-110.
Schipper, K. (1988) 'Commentary on Earnings Management', Accounting Horizons,
Vol. 3, pp. 91-102.
Schwed, F. (1940) Where are the Customers' Yachts - Or a Good Hard Look at Wall
Street, United States of America: Simon and Schuster.
Sebahattin, D. and Harlan, P. (2009) 'Financial Status, Corporate Governance Quality
and the Likelihood of Managers Using Discretionary Accruals', Accounting Research
Journal, Vol. 22, No. 2, pp. 93-117.
References
___________________________________________________________________________
93
Siagian, F. (2002) Earnings manipulation and the association between CEO bonus
and accounting earnings, Ph.D. Dissertation, University of Oregon.
Sloan, R. G. (1996) 'Do Stock Prices Fully Reflect Information in Accruals and
Cashflows about Future Earnings?', The Accounting Review, Vol. 71, No. 3,
pp. 289-315.
Spence, M. (1973) 'Job Market Signaling', The Quarterly Journal of Economics,
Vol. 87, No. 3, pp. 355-374.
Srivastava, A. (2008) Do firms use flexibility in revenue recognition rules to convey
value relevant information or to manage earnings? Working paper, Northwestern
University, Illinois, United States of America.
Stlowy, H. and Breton, G. (2004) 'Accounts Manipulation: A Literature Review and
Proposed Conceptual Framework', Review of Accounting and Finance, Vol. 3, No. 1,
pp. 5-92.
Sweeney, A. P. (1994) 'Debt covenant violations and managers' accounting
responses', Journal of Accounting and Economics, Vol. 17, No. 3, pp. 281-308.
Teoh, S., Wong, T. J. and Rao, G. (1994) ‘Incentives and Opportunities for Earnings
Management in Initial Public Offerings’ [Online]. Available at: http://deepblue.lib.
umich.edu/bitstream/2027.42/36198/2/b1738124.0001.001.pdf (Accessed 04 February
2012).
Teoh, S., Wong, T. J. and Welch, I. (1998) 'Earnings Management and the Long Run
Performance of Initial Public Offerings', The Journal of Finance, Vol. 53, No. 6,
pp. 1935-1974.
Thomson One Banker (2012) 'Thomson ONE Online Database' [Online]. Available at:
http://www.thomsonone.com/ (Accessed 30 May 2012).
References
___________________________________________________________________________
94
United States Department of the Treasury (2008) ‘Blueprint for a Modernized
Financial Regulatory Structure’ [Online]. Available at: http://www.treasury.
gov/press-center/press-releases/Documents/Blueprint.pdf. (Accessed 11 July 2012).
Van Der Wijst, N. and Hol, S. (2002) ‘Trade Credit in Europe’, Paper presented at
the 31st meeting of the European Working Group on Financial Modelling,
7-9 November 2002, Agia Napa, Cyprus.
Watts, R. L. and Zimmerman, J. L. (1986) Positive Accounting Theory, Prentice Hall.
Watts, R. L. and Zimmerman, J. L. (1978) ‘Towards a Positive Theory of the
Determination of Accounting Standards’, The Accounting Review, Vol. 53, No. 1,
pp. 112-134.
Webster, A. L. (1995) Applied Statistics for Business and Economics, 2nd Edn.
United States of America: Irwin Professional Publishing.
Wright, C. and Guan, L. (2004) 'Corporate Control and Earnings Management:
Evidence from MBOs.', Managerial Finance, Vol. 30, No. 11, pp. 45-62.
Yun, J. K. and Kim, J. J. (2011) 'Prompters of Earnings Restatement: An Exploratory
Analysis of Discretionary Accruals before and after SOX', Journal of Global Business
Management, Vol. 7, No. 2, pp. 1-6.
Zhang, Y. (2006) ‘An Empirical Analysis of Revenue Manipulation’ [Online].
Available at: http://nd.edu/~carecob/Workshops/05-06%20Workshops/Zhang.pdf
(Accessed 07 December 2011).
Appendix A
Personal
Reflection
Appendix A
___________________________________________________________________________
96
PERSONAL REFLECTION
The completion of this dissertation has left me with a great sense of achievement,
marking the high point of my experience on the MBS in Accounting programme.
As a student studying accounting over these past four years, I have been perplexed at
the extent of global corporate fraud and manipulation, often achieved through
accounting abnormalities, despite an environment of ever increasing accounting and
corporate regulation. In deciding upon a topic for my dissertation, my interest was
therefore immediately drawn towards the area of accounting manipulation and
earnings management.
After conducting preliminary research, I was surprised to find that a significant
majority of earnings management research to date has been undertaken in a U.S.
GAAP compliant financial reporting context. Having identified a literature gap with
regard to earnings management research in an IFRS compliant financial reporting
context, my attention was subsequently drawn to the methods through which
companies may manage their earnings. Moreover, having a strong personal interest in
the impact of the on-going macroeconomic uncertainty on accounting policy choice
within firms, I felt that it was important to link potential earnings management
activity with the difficulties that firms are currently experiencing.
Applying this logic, I initially considered deferred income recognition as a method of
earnings management and this formed the basis for my dissertation proposal.
However, the current limited disclosure on deferred income recognition rendered this
approach impossible. While discretionary accruals have been examined on numerous
occasions previously, I remained conscious of maintaining both the originality and
relevance of the dissertation and therefore examined individual accounting
mechanisms for earnings management. Thereafter, my focus centred upon
provisioning activity and the provision for credit loss on trade receivables more
specifically, given that credit risk has remained persistently elevated post the 2008
financial crisis. With only limited prior research in this area of earnings management,
it became clear to me that the research approach adopted would require significant
planning and articulation. However, the identifiable literature gap, originality and
relevance of the study spurred me onwards.
Appendix A
___________________________________________________________________________
97
At the time of drafting the dissertation proposal, I envisaged a timeline that included
the data collection process being substantially complete by the end of May.
While this timeline provided a tangible framework to word towards, it quickly
became evident that it was impractical, where on-going study and assignment
pressures consumed the majority of my time. I also materially underestimated the
time consumed on the data collection and analysis process, initially planning for three
weeks; with the actual process taking over 350 hours. However, the challenge of
dealing with these pressures simultaneously has benefited me greatly, instilling the
importance of rational planning while also preparing me for work in practice.
The collection of journal articles and additional literature underlying this study has
also heighted my awareness of the vast theoretical resources available that provide
insights into both earnings management and accounting choice. In order to prioritise
and succinctly summarise these resources, I developed an electronic filing system,
where each item of literature was scored relative to its applicability to the study. I feel
that this system will be of great benefit to me into the future, given the vast array of
data that is managed and processed within the accounting profession.
On reflection, I found the data collection phase of the study to be the most difficult,
where occasionally; it was difficult to appreciate any significant progress despite
many hours of work. As the months of both June and July progressed, the dissertation
was the sole focus of my daily activities, while the time remaining for completion
seemed to diminish rapidly. In retrospect, it would have been more optimal to take
greater rest breaks during this period, as the occasional sense of lethargic progress
was heavily exacerbated by exhaustion.
In completing this dissertation, I have greatly developed my skillset, having
successfully articulated and executed a large project, which enhances my confidence
as I prepare to enter the world of work. The dissertation has provided me with a great
opportunity to enhance my analytical and research capabilities, two qualities that are
fundamentally important in the audit profession, while the subject of the dissertation
is directly associated with the work of an auditor. Finally, having gained practical
experience of statistical analysis, I now appreciate, more than ever before, its
importance in explaining the rationale for the qualitative decisions undertaken that
significantly influence resultant financial performance.
Appendix B
IFRS 7 and IAS 39:
Disclosure Requirements of IFRS 7 and
Accounting Standard Guidance of IAS 39
Appendix B
___________________________________________________________________________
99
B.1 Introduction
The purpose of this appendix is to provide an overview of the disclosure requirements
of IFRS 7 and the accounting standard guidance contained in IAS 39, relating to the
provision for credit loss on trade receivables specifically.
B.2.1 Allowance for Credit Losses: IFRS 7.16 – Disclosure Requirements
“When financial assets are impaired by credit losses and the entity records the
impairment in a separate account (eg. an allowance account used to record individual
impairments or a similar account used to record a collective impairment of assets)
rather than directly reducing the carrying amount of the asset, it shall disclose a
reconciliation of changes in that account during the period for each class of financial
assets”.
B.2.2 Financial Assets that are either Past Due or Impaired: IFRS 7.37
“An entity shall disclose by class of financial asset:
a) An analysis of the age of financial assets that are past due as at the end of the
reporting period but not impaired; and
b) An analysis of financial assets that are individually determined to be impaired as
at the end of the reporting period, including the factors the entity considered in
determining that they are impaired”.
B.2.3 Impairment of Financial Assets: IAS 39.58
Section 58 of IAS 39 prescribes that:
“An entity shall assess at the end of each reporting period whether there is any
objective evidence that a financial asset or group of financial assets measured at
amortised cost is impaired. If any such evidence exists, the entity shall apply
paragraph 63 to determine the amount of any impairment loss”.
(International Accounting Standards Board, 2011)
Appendix B
___________________________________________________________________________
100
B.2.4 Impairment of Financial Assets: IAS 39.59
In determining the existence of impairment, Section 59 of IAS 39 prescribes that:
“A financial asset or a group of financial assets is impaired and impairment losses are
incurred if, and only if, there is objective evidence of impairment as a result of one or
more events that occurred after the initial recognition of the asset (a ‘loss event’) and
that loss event (or events) has an impact on the estimated future cash flows of the
financial asset or group of financial assets that can be reliability estimated. It may not
be possible to identify a single, discrete event that caused the impairment. Rather the
combined effect of several events may have caused the impairment. Losses expected
as a result of future events, no matter how likely, are not recognised. Objective
evidence that a financial asset or group of assets is impaired includes observable data
that comes to the attention of the holder of the asset about the following loss events:
Significant financial difficulty of the issuer or obligor;
A breach of contract, such as default or delinquency in interest or principal
payments;
The lender, for economic or legal reasons relating to the borrower’s financial
difficulty, granting to the borrower a concession that the lender would not
otherwise consider;
It becoming probable that the borrower will enter bankruptcy or other financial
reorganisation;
The disappearance of an active market for that financial asset because of financial
difficulties;
Observable data indicating that there is a measurable decrease in the estimated
future cash flows from a group of financial assets since the initial recognition of
those assets, although the decrease cannot yet be identified with the individual
financial assets in the group, including:
o Adverse changes in the payment status of borrowers in the group or
national or local economic conditions that correlate with defaults on the
assets in the group”.
Appendix B
___________________________________________________________________________
101
B.2.5 Impairment of Financial Assets: IAS 39.63
In determining the amount of impairment loss, Section 63 of IAS 39 prescribes that:
“If there is objective evidence that an impairment loss on financial assets measured at
amortised cost has been incurred, the amount of the loss is measured as the difference
between the asset’s carrying amount and the present value of estimated future cash
flows (excluding future credit losses that have not been incurred) discounted at the
financial asset’s original effective interest rate (i.e. the effective interest rate
computed at initial recognition). The carrying amount of the asset shall be reduced
either directly or through the use of an allowance account. The amount of the loss
shall be recognised in the profit or loss”.
(International Accounting Standards Board, 2011)
Appendix C
Details of Final
Sample Population:
Company Details and Sector Specific
Composition of Final Sample Population
Appendix C
___________________________________________________________________________
103
C.1 Introduction
The purpose of this appendix is to provide detail regarding the names and sector
specific composition of the final sample of 204 FTSE 350 companies selected for
analysis in this study. Specific details relating to each individual company are
presented, along with a sector specific composition summary table.
COMPANY NAME FTSE 350 SECTOR
AEGIS GROUP PLC Media
AGGREKO PLC Support Services
AMEC PLC Oil Equipment and Services
ANGLO AMERICAN PLC Mining
ANTOFAGASTA PLC Mining
ARM HOLDINGS PLC Technology and Hardware Equipment
ASHTEAD GROUP PLC Support Services
ASSOCIATED BRITISH FOODS PLC Food Producers
ASTRAZENECA PLC Pharmaceuticals and Biotechnology
AVEVA GROUP PLC Software and Computer Services
AZ ELECTRONIC MATERIALS PLC Chemicals
BABCOCK INTERNATIONAL GROUP PLC Support Services
BAE SYSTEMS PLC Aerospace and Defense
BALFOUR BEATTY PLC Construction and Materials
BARR A.G. PLC Beverages
BARRATT DEVELOPMENTS PLC Household Goods and Home Construction
BBA AVIATION PLC Industrial Transportation
BERENDSEN PLC Support Services
BERKELEY GROUP HOLDINGS PLC Household Goods and Home Construction
BG GROUP PLC Oil and Gas Producers
BHP BILLITON PLC Mining
BODYCOTE PLC Industrial Engineering
BOOKER GROUP PLC Food and Drug Retailers
BOVIS HOMES GROUP PLC Household Goods and Home Construction
BP PLC Oil and Gas Producers
BRITVIC PLC Beverages
BSKYB GROUP PLC Media
BT GROUP PLC Fixed Line Telecommunications
BTG PLC Pharmaceuticals and Biotechnology
BUMI PLC Mining
Appendix C
___________________________________________________________________________
104
COMPANY NAME FTSE 350 SECTOR
BUNZL PLC Support Services
BURBERRY GROUP PLC Personal Goods
BWIN.PARTY DIGITAL ENTER. PLC Travel and Leisure
CABLE & WIRELESS COMM PLC Fixed Line Telecommunications
CABLE & WIRELESS W.W. PLC Fixed Line Telecommunications
CAPE PLC Oil Equipment and Services
CAPITA GROUP PLC Support Services
CARILLION PLC Support Services
CARPETRIGHT PLC General Retailers
CENTRICA PLC Gas, Water and Multiutilities
CHEMRING GROUP PLC Aerospace and Defense
COBHAM PLC Aerospace and Defense
COLT GROUP S.A. Fixed Line Telecommunications
COMPASS GROUP PLC Travel and Leisure
COMPUTACENTER PLC Software and Computer Services
COOKSON GROUP PLC General Industrials
CRANSWICK PLC Food Producers
CRH PLC Construction and Materials
CRODA INTERNATIONAL PLC Chemicals
CSR PLC Technology and Hardware Equipment
DAILY MAIL & GEN TRUST PLC Media
DAIRY CREST GRP PLC Food Producers
DE LA RUE PLC Support Services
DEBENHAMS PLC General Retailers
DEVRO PLC Food Producers
DIAGEO PLC Beverages
DIGNITY PLC General Retailers
DIPLOMA PLC Support Services
DIXONS RETAIL PLC General Retailers
DOMINO PRINTING PLC Electronic and Electrical Equipment
DOMINO'S PIZZA GROUP PLC Travel and Leisure
DRAX GROUP PLC Electricity
DS SMITH PLC General Industrials
EASYJET PLC Travel and Leisure
ELECTROCOMPONENTS PLC Support Services
Appendix C
___________________________________________________________________________
105
COMPANY NAME FTSE 350 SECTOR
ELEMENTIS PLC Chemicals
ENQUEST PLC Oil and Gas Producers
ESSAR ENERGY PLC Oil and Gas Producers
EURAS. NATURAL RES. CORP PLC Mining
EUROMONEY INST INVESTOR PLC Media
EVRAZ PLC Industrial Metals and Mining
EXILLON ENERGY PLC Oil and Gas Producers
EXPERIAN PLC Support Services
FENNER PLC Industrial Engineering
FERREXPO PLC Industrial Metals and Mining
FIDESSA GROUP PLC Software and Computer Services
FIRSTGROUP PLC Travel and Leisure
FRESNILLO PLC Mining
G4S PLC Support Services
GALLIFORD TRY PLC Construction and Materials
GENUS PLC Pharmaceuticals and Biotechnology
GKN PLC Automobiles and Parts
GLAXOSMITHKLINE PLC Pharmaceuticals and Biotechnology
GLENCORE INTERNATIONAL PLC Mining
GO-AHEAD GROUP PLC Travel and Leisure
GREENE KING PLC Travel and Leisure
HALFORDS GROUP PLC General Retailers
HALMA PLC Electronic and Electrical Equipment
HAYS PLC Support Services
HIKMA PHARMACEUTICALS PLC Pharmaceuticals and Biotechnology
HOME RETAIL GROUP PLC General Retailers
HOMESERVE PLC Support Services
HOWDEN JOINERY GROUP PLC Support Services
HUNTING PLC Oil Equipment and Services
IMAGINATION TECH. GROUP PLC Technology and Hardware Equipment
IMI PLC Industrial Engineering
IMPERIAL TOBACCO GROUP PLC Tobacco
INCHCAPE PLC General Retailers
INFORMA PLC Media
INMARSAT PLC Mobile Telecommunications
Appendix C
___________________________________________________________________________
106
COMPANY NAME FTSE 350 SECTOR
INTERCONTINENTAL HOTELS GRP. PLC Travel and Leisure
INTERSERVE PLC Support Services
INTERTEK GROUP PLC Support Services
INVENSYS PLC Software and Computer Services
ITE GROUP PLC Media
JD SPORTS FASHION PLC General Retailers
JOHN WOOD GROUP PLC Oil Equipment and Services
JOHSON MATTHEY PLC Chemicals
KAZAKHMYS PLC Mining
KCOM GROUP PLC Fixed Line Telecommunications
KESA ELECTRICALS PLC General Retailers
KIER GROUP PLC Construction and Materials
KINGFISHER PLC General Retailers
LADBROKES PLC Travel and Leisure
LAIRD PLC Technology and Hardware Equipment
LAMPRELL PLC Oil Equipment and Services
LOGICA PLC Software and Computer Services
LONMIN PLC Mining
MARKS & SPENCER GROUP PLC General Retailers
MARSTON'S PLC Travel and Leisure
MEGGITT PLC Aerospace and Defense
MELROSE PLC Industrial Engineering
MICHAEL PAGE INTERNATIONAL PLC Support Services
MICRO FOCUS INTERNATIONAL PLC Software and Computer Services
MILLENNIUM AND COP. HOTELS PLC Travel and Leisure
MITIE GROUP PLC Support Services
MONDI GROUP PLC Forestry and Paper
MONEYSUPERMARKET.COM PLC Media
MORGAN CRUCIBLE PLC Electronic and Electrical Equipment
MORRISON SUPERMARKETS PLC Food and Drug Retailers
N BROWN GROUP PLC General Retailers
NATIONAL EXPRESS GROUP PLC Travel and Leisure
NATIONAL GRID PLC Gas, Water and Multiutilities
NEXT PLC General Retailers
NORTHGATE PLC Support Services
Appendix C
___________________________________________________________________________
107
COMPANY NAME FTSE 350 SECTOR
OCADO GROUP PLC Food and Drug Retailers
OXFORD INSTRUMENTS PLC Electronic and Electrical Equipment
PAYPOINT PLC Support Services
PEARSON PLC Media
PENNON GROUP PLC Gas, Water and Multiutilities
PERFORM GROUP PLC Media
PERSIMMON PLC Household Goods and Home Construction
PETROFAC LIMITED Oil Equipment and Services
PETROPAVLOVSK PLC Mining
POLYMETAL INTERNATIONAL PLC Mining
PREMIER FARNELL PLC Support Services
PZ CUSSONS PLC Personal Goods
QINETIQ PLC Aerospace and Defense
RANDGOLD RESOURCES LTD Mining
RECKITT BENCKISER GROUP PLC Household Goods and Home Construction
REED ELSEVIER PLC Media
REGUS PLC Support Services
RENISHAW PLC Electronic and Electrical Equipment
RENTOKIL INITIAL PLC Support Services
REXAM PLC General Industrials
RIGHTMOVE PLC Media
RIO TINTO PLC Mining
ROTORK PLC Industrial Engineering
ROYAL DUTCH SHELL PLC Oil and Gas Producers
RPC GROUP PLC General Industrials
RPS GROUP PLC Support Services
SABMILLER PLC Beverages
SAGE GROUP PLC Software and Computer Services
SALAMANDER ENERGY PLC Oil and Gas Producers
SDL PLC Software and Computer Services
SENIOR PLC Aerospace and Defense
SERCO GROUP PLC Support Services
SEVERN TRENT PLC Gas, Water and Multiutilities
SHANKS GROUP PLC Support Services
SHIRE PLC Pharmaceuticals and Biotechnology
Appendix C
___________________________________________________________________________
108
COMPANY NAME FTSE 350 SECTOR
SIG PLC Support Services
SMITH & NEPHEW PLC Health Care Equipment and Services
SMITHS GROUP PLC General Industrials
SPECTRIS PLC Electronic and Electrical Equipment
SPIRAX-SARCO ENGINEERING PLC Industrial Engineering
SPIRENT COMMUNICATIONS PLC Technology and Hardware Equipment
SPORTS DIRECT INTERNATIONAL PLC General Retailers
SSE PLC Electricity
STAGECOACH GROUP PLC Travel and Leisure
STOBART GROUP LTD Industrial Transportation
SUPERGROUP PLC Personal Goods
SYNERGY HEALTH PLC Health Care Equipment and Services
TALKTALK PLC Fixed Line Telecommunications
TALVIVAARA MINING CO. PLC Industrial Metals and Mining
TATE & LYLE PLC Food Producers
TELECITY GROUP PLC Software and Computer Services
TESCO PLC Food and Drug Retailers
TRAVIS PERKINS PLC Support Services
TUI TRAVEL PLC Travel and Leisure
TULLOW OIL PLC Oil and Gas Producers
UBM PLC Media
ULTRA ELECTRONICS PLC Aerospace and Defense
UNILEVER PLC Food Producers
UNITED UTILITIES GROUP PLC Gas, Water and Multiutilities
VICTREX PLC Chemicals
VODAFONE GROUP PLC Mobile Telecommunications
WEIR GROUP PLC Industrial Engineering
WH SMITH PLC General Retailers
WHITBREAD PLC Travel and Leisure
WOLSELEY PLC Support Services
WPP PLC Media
WS ATKINS PLC Support Services
XSTRATA PLC Mining
YULE CATTO & CO. PLC Chemicals
Note: A sector specific composition summary table is provided overleaf.
Appendix C
___________________________________________________________________________
109
Figure C.1 - FTSE 350 Sector Specific Composition Summary
FTSE 350 SECTOR SPECIFIC COMPOSITION SUMMARY
FTSE 350 - Sector N %
Aerospace and Defense 7 3.4%
Automobiles and Parts 1 0.5%
Beverages 4 2.0%
Chemicals 6 2.9%
Construction and Materials 4 2.0%
Electricity 2 1.0%
Electronic and Electrical Equipment 6 2.9%
Fixed Line Telecommunications 6 2.9%
Food and Drug Retailers 4 2.0%
Food Producers 6 2.9%
Forestry and Paper 1 0.5%
Gas, Water and Multiutilities 5 2.5%
General Industrials 5 2.5%
General Retailers 15 7.4%
Health Care Equipment and Services 2 1.0%
Household Goods and Home Construction 5 2.5%
Industrial Engineering 7 3.4%
Industrial Metals and Mining 3 1.5%
Industrial Transportation 2 1.0%
Media 13 6.4%
Mining 14 6.9%
Mobile Telecommunications 2 1.0%
Oil and Gas Producers 8 3.9%
Oil Equipment and Services 6 2.9%
Personal Goods 3 1.5%
Pharmaceuticals and Biotechnology 6 2.9%
Software and Computer Services 9 4.4%
Support Services 31 15.2%
Technology and Hardware Equipment 5 2.5%
Tobacco 1 0.5%
Travel & Leisure 15 7.4%
Total 204
Appendix D
Multiple Regression
Analysis Data:
Multiple Regression Analysis Results:
Outliers, Secondary Dependent Variable and
Stock Price Performance
Appendix D
___________________________________________________________________________
111
D.1 Introduction
The impact of the six identified outliers upon preliminary testing along with the
results of multiple regression analysis that examine both the secondary dependent
variable and the significance of post financial year end stock price performance in the
study are detailed in this appendix.
D.2. Impact of Outliers upon Preliminary Testing
In analysing the determinants of earnings management (abnormal provision), removal
of the six identified outliers from preliminary regression analysis resulted in a marked
difference in the significance statistics of Multiple Regression One (Table D.1 below).
Table D.1 – Impact of Outliers upon Preliminary Regression Analysis
Identified Outliers R Square Adj. R
Square F Stat
P Value
(F Stat)
Inclusion of Six Outliers 0.841 0.829 73.739 0.000
Exclusion of Six Outliers 0.083 0.015 1.215 0.266
Although the significantly positive R Square and Adj. R Square values of 0.841 and
0.829 indicate apparently strong explanatory power, the removal of Heritage Oil PLC
resulted in a 45% reduction in the Adj. R Square value, indicating the impact of such
outliers on the analysis. All six outliers were therefore excluded from further testing.
D.3 Secondary Dependent Variable
Measure of Abnormal Provision – Proxy for Earnings Management Activity
The secondary dependent variable comprises the relative change in the provision for
credit loss on trade receivables, without controlling for changes in total gross trade
receivables.
{ } :
Where: EQ = Earnings quality.
Prov. (t) = Provision for credit loss on trade receivables in latest financial period.
Prov. (t-1) = Provision for credit loss on trade receivables in previous financial period.
Prov. (t) - Prov. (t-1)
Prov. (t-1)
X 100 EQ
Appendix D
___________________________________________________________________________
112
D.3.1 Alternative Regression One: Fifteen Hypotheses Regression (N=204)
Alternative multiple regression one consists of the full sample of 204 FTSE 350
companies and tests all fifteen hypotheses as indicated in Figure D.1 below, where the
relative change in total gross trade receivables is added as a fifteenth independent
explanatory variable.
Figure D.1 – Alternative Multiple Regression One Equation
The title of each variable in the equation is shortened in the interest of brevity.
The resultant regression significance statistics and fifteen hypotheses results are
contained in Tables D.2 and D.3 below and overleaf.
Table D.2 – Alternative Multiple Regression One Significance Statistics (N=204)
β0 Constant R Square 0.186
Coefficient -0.367 Adj. R Square 0.121
T-Stat -0.751 F Stat 2.859
P-Value 0.453 P-Value F Stat 0.000
The results contained in Table D.2 above indicate that the model has explanatory
power at all levels of significance, with a regression significance P-Value of 0.000.
The Adj. R Square value of 0.121 indicates that the model explains 12.1 per cent of
the variation in the change in provision for credit loss on trade receivables. The results
for all 15 hypotheses tested are contained in Table D.3 overleaf.
Earnings Quality = β0 + β1 EPS GROWTH+ β2 EPS SURPRISE + β3 BONUS
+ β4 REMUN + β5 GEARING + β6 GROSS MARGIN
+ β7 NET MARGIN + β8 T/REC DAYS + β9 INEDS BOARD
+ β10 INEDS AUDIT + β11 GOV NON COMP
+ β12 AUDIT MEET + β13 AUDITOR + β14 AUDIT FEES
+ β15 GROSS T/REC + 𝜀
Appendix D
___________________________________________________________________________
113
Table D.3 - Alternative Multiple Regression One: Fifteen Hypotheses (N=204)
Hypo. Variable Pred.
Sign Coefficient P-Value
2 β1 Consensus EPS Growth (%) - -0.005 0.732
3 β2 Earnings (EPS) Surprise (%) - 0.119 0.319
4 β3 Existence of Bonus Plan - -0.086 0.714
5 β4 Executive Incentive Remuneration - -0.062 0.294
6 β5 Change in Gearing (%) - -0.001 0.835
7 β6 Change in Gross Margin (%) - -0.004 0.259
8 β7 Change in Net Margin (%) - -0.002 0.446
9 β8 Change in Trade Rec. Days - 0.000 0.805
10 β9 Proportion of INEDs to the Total Board + 0.144 0.545
11 β10 Proportion of INEDs to Audit Committee + 0.759 0.045*
12 β11 # Of Governance Non Compliance Issues - 0.035 0.165
13 β12 # Of Audit Committee Meetings + -0.008 0.585
14 β13 Auditor Type + -0.377 0.008*
15 β14 Audit Specific Fee(s) + 0.287 0.480
1 β15 Change in Gross Trade Receivables + 0.243 0.004
Bolded P-Values in the above Table indicate statistical significance at the 1% level.
One variable: the Change in Gross Trade Receivables exhibits a significant positive
relationship with the change in provision for credit loss on trade receivables.
While the Auditor Type (0.008*) and Proportion of INEDs to the Audit Committee
(0.045*) variables exhibit apparent significance at the 1% and 5% levels respectively,
these results are deemed to be skewed and are disregarded, given that both variables
are not normally distributed, where only six of the 204 FTSE 350 companies were
audited by a non-Big 4 auditor and where only eight of the total sample reported audit
committee structures that comprised of less than 100% INED composition. The
remaining variables exhibit no significant relationship with the change in provision
for credit loss on trade receivables.
Appendix D
___________________________________________________________________________
114
D.4 Multiple Regression Analysis: Stock Price Performance
Regression One: Extreme Abnormal Underprovision (N=25)
Regression one consists of the full sample of 25 FTSE 350 extreme abnormal
underproviders and comprises the variables in Figure D.2 below.
Figure D.2 – Stock Price Performance Multiple Regression One Equation
The title of each variable in the equation is shortened in the interest of brevity.
Where:
Stock Performance = Stock price performance post financial year end (Section 5.9.2).
β0 = Intercept and 𝜀 = Regression error term.
β1 = Extreme abnormal underprovision for credit loss on trade receivables.
β2 = Stock beta value (Risk). β3 = Natural Logarithm of Total Assets (Size).
Table D.4 – Stock Price Performance Regression One Sig. Statistics (N=25)
β0 Constant R Square 0.219
Coefficient 0.164 Adj. R Square 0.107
T-Stat 0.462 F Stat 1.958
P-Value 0.649 P-Value F Stat 0.151
The results contained in Table D.4 indicate explanatory power, with an Adj. R Square
value of 0.107 indicating that the model explains 10.7 per cent of the variation in the
stock price performance. However, a regression significance P-Value of 0.151
indicates explanatory insignificance at all levels. The associated hypotheses results
are contained in Table D.5 overleaf.
Stock Performance = β0 + β1 EARNINGS QUALITY+ β2 BETA VALUE
+ β3 NAT. LOG ASSETS (SIZE) + 𝜀
Appendix D
___________________________________________________________________________
115
Table D.5 – Regression One: Stock Price Performance Hypotheses (N=25)
Variable Coefficient P-Value
β1 Ext. Abnormal U/Provision for Credit Loss 0.405 0.032
β2 Stock Beta Value (Risk) 0.035 0.721
β3 Nat. Log of Total Assets (Size) -0.002 0.985
Bolded P-Values in the above Table indicate statistical significance at the 5% level.
The insignificance of the Stock Beta Value and Natural Logarithm of Total Assets
variables is most surprising (Table D.5 above), as both variables normally display a
statistically significant relationship with resultant stock price performance.
Notably, the extent of extreme abnormal underprovision for credit loss on trade
receivables exhibits a significant positive (5% level) relationship with resultant stock
price performance. Regression one was therefore re-run, with the exclusion of β2 and
β3 given their insignificance.
Figure D.3 – Stock Price Performance Regression Two Equation
The title of each variable in the equation is shortened in the interest of brevity.
Table D.6 – Stock Price Performance Regression Two Sig. Statistics (N=25)
β0 Constant R Square 0.215
Coefficient 0.199 Adj. R Square 0.179
T-Stat 1.482 F Stat 6.249
P-Value 0.151 P-Value F Stat 0.020
A marked increase in the F Stat (6.249) and regression significance P-Value (0.020)
are noted, while the Adj. R Square value of 0.179 indicates that the model explains
17.9 per cent of the variation in the stock price performance.
Table D.7 – Regression Two: Stock Price Performance Hypothesis (N=25)
Variable Coefficient P-Value
β1 Ext. Abnormal U/Provision for Credit Loss 0.416 0.020
Bolded P-Values in the above Table indicate statistical significance at the 5% level.
Stock Performance = β0 + β1 EARNINGS QUALITY+ 𝜀
Appendix D
___________________________________________________________________________
116
The extent of extreme abnormal underprovision for credit loss on trade receivables
exhibits a significant positive (5% level) relationship with resultant stock price
performance, where a 1% increase in earnings quality (reduction in the extent of
abnormal underprovision) results in a 0.416% more positive stock price performance.
Regression Three: Extreme Abnormal Overprovision (N=25)
Regression three consists of the full sample of 25 FTSE 350 extreme abnormal
overproviders and comprises the variables in Figure D.4 below.
Figure D.4 – Stock Price Performance Multiple Regression Three Equation
The title of each variable in the equation is shortened in the interest of brevity.
Where:
Stock Performance = Stock price performance post financial year end (Section 5.9.2).
β0 = Intercept and 𝜀 = Regression error term.
β1 = Extreme abnormal overprovision for credit loss on trade receivables.
β2 = Stock beta value (Risk). β3 = Natural Logarithm of Total Assets (Size).
Table D.8 – Stock Price Performance Regression Three Sig. Statistics (N=25)
β0 Constant R Square 0.259
Coefficient 0.668 Adj. R Square 0.153
T-Stat 2.799 F Stat 2.445
P-Value 0.010 P-Value F Stat 0.092
The Adj. R Square value (0.153) indicates that the model explains 15.3 per cent of the
variation in the stock price performance, with explanatory significance at the 10%
level.
Stock Performance = β0 + β1 EARNINGS QUALITY+ β2 BETA VALUE
+ β3 NAT. LOG ASSETS (SIZE) + 𝜀
Appendix D
___________________________________________________________________________
117
Table D.9 – Regression Three: Stock Price Performance Hypotheses (N=25)
Variable Coefficient P-Value
β1 Ext. Abnormal O/Provision for Credit Loss -0.161 0.328
β2 Stock Beta Value (Risk) -0.456 0.016
β3 Nat. Log Total Assets (Size) 0.968 0.020
Bolded P-Values in the above Table indicate statistical significance at the 5% level.
The significance of the Stock Beta Value (5% level) and Natural Logarithm of Total
Assets (5% level) variables is evident in Table D.9 above, with a significant negative
relationship between the Stock Beta Value variable and resultant stock price
performance and a significant positive relationship in the case of the Natural
Logarithm of Total Assets variable. Notably, the extent of abnormal overprovision for
credit loss on trade receivables exhibits no significant relationship with resultant stock
price performance.
Appendix E
Disclosure Notes:
Disclosure Extracts from Annual Reports:
Trade Receivables
Appendix E
___________________________________________________________________________
119
E.1 Introduction
The purpose of this appendix is to provide extracts of the primary disclosure notes
from the financial statements that were utilised in order to gather data in this study.
Example One: QINETIC PLC – Trade and Other Receivables
Example Two: QINETIC PLC – Credit Risk Disclosure Note
While example one details the annual movement on trade receivables and the
associated provision, example two details information regarding the entity’s
assessment of credit risk. This information is utilised to determine the extent of
abnormal provision for credit loss on trade receivables during the latest period.
Appendix F
Dependent Variable
Dataset:
Company Specific
Dependent Variable Data
Appendix F
___________________________________________________________________________
121
F.1 Introduction
The dataset relating to both the primary and secondary dependent variables utilised in
this study, for the final sample of 204 FTSE 350 companies, is contained in this
appendix.
Primary Dependent Variable
The primary dependent variable comprises the relative change in the provision for
credit loss on trade receivables after controlling for the relative change in total gross
trade receivables, which is calculated as follows:
{ }-{ }: Secondary Dependent Variable
The secondary dependent variable comprises the relative change in the provision for
credit loss on trade receivables, without controlling for changes in total gross trade
receivables.
{ } : COMPANY NAME
Primary Dependent
Variable
Secondary
Dependent Variable
AEGIS GROUP PLC -0.96% -1.03%
AGGREKO PLC -24.23% 9.09%
AMEC PLC -41.18% -9.09%
ANGLO AMERICAN PLC 5.86% 1.89%
ANTOFAGASTA PLC 13.48% 14.71%
ARM HOLDINGS PLC -38.62% -21.14%
ASHTEAD GROUP PLC -24.58% -12.18%
ASSOCIATED BRITISH FOODS PLC -11.91% 0.00%
ASTRAZENECA PLC -24.33% -18.52%
AVEVA GROUP PLC -96.49% -45.04%
Prov. (t) - Prov. (t-1)
Prov. (t-1)
X 100
GTR (t) - GTR (t-1)
GTR (t-1)
X 100 EQ
Prov. (t) - Prov. (t-1)
Prov. (t-1)
X 100 EQ
Appendix F
___________________________________________________________________________
122
COMPANY NAME Primary Dependent
Variable
Secondary
Dependent Variable
AZ ELECTRONIC MATERIALS PLC -42.31% -45.00%
BABCOCK INTERNATIONAL GROUP PLC -14.01% -35.35%
BAE SYSTEMS PLC 3.86% -4.88%
BALFOUR BEATTY PLC -22.05% -4.17%
BARR A.G. PLC 42.25% 58.58%
BARRATT DEVELOPMENTS PLC 27.50% 26.67%
BBA AVIATION PLC -9.30% -6.06%
BERENDSEN PLC -5.69% -12.07%
BERKELEY GROUP HOLDINGS PLC -85.38% 0.00%
BG GROUP PLC -23.75% 7.45%
BHP BILLITON PLC -19.41% 2.72%
BODYCOTE PLC -16.65% -10.14%
BOOKER GROUP PLC -19.30% -8.51%
BOVIS HOMES GROUP PLC -141.60% 0.00%
BP PLC -35.85% -22.43%
BRITVIC PLC -13.31% 0.00%
BSKYB GROUP PLC 37.02% 27.45%
BT GROUP PLC -1.18% -2.60%
BTG PLC -22.83% -22.22%
BUMI PLC -100.00% 0.00%
BUNZL PLC -18.11% -11.54%
BURBERRY GROUP PLC -39.47% -37.19%
BWIN.PARTY DIGITAL ENTER. PLC -56.42% 107.14%
CABLE & WIRELESS COMM PLC 3.37% 8.06%
CABLE & WIRELESS W.W. PLC -22.21% -30.30%
CAPE PLC -87.33% -64.00%
CAPITA GROUP PLC 11.46% 32.60%
CARILLION PLC 31.33% 32.05%
CARPETRIGHT PLC -4.94% 0.00%
CENTRICA PLC -10.11% -6.55%
CHEMRING GROUP PLC -47.29% -30.00%
COBHAM PLC -6.75% -16.67%
COLT GROUP S.A. 18.88% 15.71%
COMPASS GROUP PLC -15.01% -5.06%
COMPUTACENTER PLC -15.14% 0.79%
Appendix F
___________________________________________________________________________
123
COMPANY NAME Primary Dependent
Variable
Secondary
Dependent Variable
COOKSON GROUP PLC -9.98% -11.50%
CRANSWICK PLC -4.60% -12.68%
CRH PLC -11.11% 1.32%
CRODA INTERNATIONAL PLC -28.20% -31.34%
CSR PLC -36.91% -22.00%
DAILY MAIL & GEN TRUST PLC -2.19% -5.15%
DAIRY CREST GRP PLC -16.34% -8.62%
DE LA RUE PLC 14.61% 20.51%
DEBENHAMS PLC 87.06% 100.00%
DEVRO PLC 3.32% 9.09%
DIAGEO PLC -0.39% 0.00%
DIGNITY PLC -0.40% -16.22%
DIPLOMA PLC -32.46% -16.67%
DIXONS RETAIL PLC -7.45% -13.57%
DOMINO PRINTING PLC -30.33% -25.98%
DOMINO'S PIZZA GROUP PLC -104.29% 0.00%
DRAX GROUP PLC 9.58% 15.38%
DS SMITH PLC -17.75% -1.13%
EASYJET PLC -9.09% 0.00%
ELECTROCOMPONENTS PLC -14.35% -14.55%
ELEMENTIS PLC 4.29% -7.14%
ENQUEST PLC 3.74% 0.00%
ESSAR ENERGY PLC -20.26% 0.00%
EURAS. NATURAL RES. CORP PLC 135.34% 154.55%
EUROMONEY INST INVESTOR PLC -21.12% -4.22%
EVRAZ PLC 8.07% -11.22%
EXILLON ENERGY PLC -13.91% 3.68%
EXPERIAN PLC -23.67% -21.28%
FENNER PLC -40.39% -11.54%
FERREXPO PLC -74.39% -61.76%
FIDESSA GROUP PLC -6.53% -1.71%
FIRSTGROUP PLC 26.96% 15.38%
FRESNILLO PLC -13.11% -9.81%
G4S PLC -13.02% -8.22%
GALLIFORD TRY PLC 4.32% -33.33%
Appendix F
___________________________________________________________________________
124
COMPANY NAME Primary Dependent
Variable
Secondary
Dependent Variable
GENUS PLC -1.12% 6.06%
GKN PLC -6.41% 20.00%
GLAXOSMITHKLINE PLC 13.98% 8.15%
GLENCORE INTERNATIONAL PLC -41.85% -16.77%
GO-AHEAD GROUP PLC -21.26% -27.78%
GREENE KING PLC -6.56% 15.69%
HALFORDS GROUP PLC -1.96% 0.00%
HALMA PLC 25.48% 37.29%
HAYS PLC -19.12% 11.11%
HIKMA PHARMACEUTICALS PLC -47.57% -17.45%
HOME RETAIL GROUP PLC 9.65% 8.18%
HOMESERVE PLC -54.99% -44.97%
HOWDEN JOINERY GROUP PLC -10.04% -5.43%
HUNTING PLC -51.22% 29.41%
IMAGINATION TECH. GROUP PLC 60.87% 150.00%
IMI PLC -4.67% 4.60%
IMPERIAL TOBACCO GROUP PLC 0.71% 0.00%
INCHCAPE PLC -9.89% -1.28%
INFORMA PLC -1.00% 6.79%
INMARSAT PLC 9.65% 0.00%
INTERCONTINENTAL HOTELS GRP. PLC -19.26% -20.69%
INTERSERVE PLC -1.02% 2.11%
INTERTEK GROUP PLC 12.93% 49.51%
INVENSYS PLC -0.66% -8.33%
ITE GROUP PLC 12.67% 32.00%
JD SPORTS FASHION PLC -11.09% 18.33%
JOHN WOOD GROUP PLC -23.30% -9.91%
JOHSON MATTHEY PLC -44.84% -4.00%
KAZAKHMYS PLC 27.07% 5.08%
KCOM GROUP PLC -17.31% -31.14%
KESA ELECTRICALS PLC -20.45% -7.89%
KIER GROUP PLC -20.43% -27.03%
KINGFISHER PLC -9.72% -25.00%
LADBROKES PLC -29.39% -33.33%
LAIRD PLC -21.72% -20.00%
Appendix F
___________________________________________________________________________
125
COMPANY NAME Primary Dependent
Variable
Secondary
Dependent Variable
LAMPRELL PLC -118.93% 3.74%
LOGICA PLC 17.61% 18.03%
LONMIN PLC 74.05% 0.00%
MARKS & SPENCER GROUP PLC -23.96% -7.69%
MARSTON'S PLC 55.94% 42.86%
MEGGITT PLC 44.76% 84.62%
MELROSE PLC -36.14% -51.52%
MICHAEL PAGE INTERNATIONAL PLC -5.38% 10.88%
MICRO FOCUS INTERNATIONAL PLC 102.76% 98.47%
MILLENNIUM AND COP. HOTELS PLC -8.26% 9.09%
MITIE GROUP PLC -15.91% -23.19%
MONDI GROUP PLC -1.63% -15.69%
MONEYSUPERMARKET.COM PLC -35.34% -37.65%
MORGAN CRUCIBLE PLC -3.30% 2.17%
MORRISON SUPERMARKETS PLC 27.49% 25.00%
N BROWN GROUP PLC 3.13% 9.31%
NATIONAL EXPRESS GROUP PLC -32.33% -28.67%
NATIONAL GRID PLC 1.02% -9.00%
NEXT PLC -6.18% 4.50%
NORTHGATE PLC 39.57% 30.04%
OCADO GROUP PLC -53.49% 15.38%
OXFORD INSTRUMENTS PLC 19.02% 0.00%
PAYPOINT PLC 44.63% 13.99%
PEARSON PLC 19.38% 22.89%
PENNON GROUP PLC 3.09% 15.02%
PERFORM GROUP PLC 96.70% 191.18%
PERSIMMON PLC 14.85% 0.00%
PETROFAC LIMITED -52.88% -42.60%
PETROPAVLOVSK PLC -14.42% 0.00%
POLYMETAL INTERNATIONAL PLC -90.28% 0.00%
PREMIER FARNELL PLC -0.68% -2.08%
PZ CUSSONS PLC -54.75% -40.00%
QINETIQ PLC -22.28% -34.09%
RANDGOLD RESOURCES LTD -133.73% 0.00%
RECKITT BENCKISER GROUP PLC 10.53% 14.63%
Appendix F
___________________________________________________________________________
126
COMPANY NAME Primary Dependent
Variable
Secondary
Dependent Variable
REED ELSEVIER PLC -13.70% -13.70%
REGUS PLC -2.71% 0.00%
RENISHAW PLC -30.99% 0.45%
RENTOKIL INITIAL PLC -6.13% -4.74%
REXAM PLC -50.60% -47.06%
RIGHTMOVE PLC -14.21% 15.63%
RIO TINTO PLC 0.09% -5.41%
ROTORK PLC -26.02% 10.78%
ROYAL DUTCH SHELL PLC -30.05% -1.45%
RPC GROUP PLC -14.96% 32.69%
RPS GROUP PLC 10.07% 25.05%
SABMILLER PLC -3.57% -5.77%
SAGE GROUP PLC -0.07% 6.57%
SALAMANDER ENERGY PLC 85.16% 0.00%
SDL PLC 6.38% 2.42%
SENIOR PLC -36.39% 0.00%
SERCO GROUP PLC 65.84% 71.43%
SEVERN TRENT PLC 7.92% 7.30%
SHANKS GROUP PLC -17.21% -2.17%
SHIRE PLC 10.08% 32.91%
SIG PLC -7.24% -9.86%
SMITH & NEPHEW PLC -27.18% -26.53%
SMITHS GROUP PLC -9.21% -4.11%
SPECTRIS PLC 16.61% 30.77%
SPIRAX-SARCO ENGINEERING PLC -14.23% -11.54%
SPIRENT COMMUNICATIONS PLC 33.23% 50.00%
SPORTS DIRECT INTERNATIONAL PLC 23.86% 30.56%
SSE PLC -34.07% -3.16%
STAGECOACH GROUP PLC -47.06% -57.78%
STOBART GROUP LTD 72.14% 58.76%
SUPERGROUP PLC -132.00% 0.00%
SYNERGY HEALTH PLC -38.69% -31.92%
TALKTALK PLC 15.66% -16.22%
TALVIVAARA MINING CO. PLC -22.30% 0.00%
TATE & LYLE PLC -0.77% -20.83%
Appendix F
___________________________________________________________________________
127
COMPANY NAME Primary Dependent
Variable
Secondary
Dependent Variable
TELECITY GROUP PLC -37.89% -33.78%
TESCO PLC -16.36% 13.64%
TRAVIS PERKINS PLC 20.96% 22.75%
TUI TRAVEL PLC 17.26% 5.45%
TULLOW OIL PLC -71.43% 0.00%
UBM PLC -4.88% 1.67%
ULTRA ELECTRONICS PLC 48.42% 81.37%
UNILEVER PLC -14.21% -0.85%
UNITED UTILITIES GROUP PLC -7.38% -73.28%
VICTREX PLC -46.31% -25.00%
VODAFONE GROUP PLC 5.79% 0.80%
WEIR GROUP PLC -53.37% -7.25%
WH SMITH PLC 47.14% 50.00%
WHITBREAD PLC 12.66% 9.09%
WOLSELEY PLC -29.24% -21.67%
WPP PLC 9.13% 9.69%
WS ATKINS PLC -43.23% -28.07%
XSTRATA PLC 22.72% 0.00%
YULE CATTO & CO. PLC 34.05% 60.35%
Appendix G
Methodology
Continued:
Additional Variables, Measures and
OLS Regression Assumptions
Appendix G
___________________________________________________________________________
129
G.1 Introduction
The purpose of this appendix is to provide additional detail relating to the
composition of the various independent variables and measures utilised in this study,
where such information is not presented in the Methodology chapter. This appendix
also details the results of various procedures undertaken to ensure compliance with
the underlying assumptions of OLS regression analysis.
G2. Explanatory Detail – Additional Variables and Measures
For all of the following variables and measures:
(t) = Latest financial period (latest available annual report financial period).
(t-1) = Previous financial period.
G.2.1 Analyst Consensus EPS Growth %
{ } G.2.2 Earnings (EPS) Surprise %
Obtained directly from the Thomson One Banker database, this variable comprises the
actual EPS surprise % for the financial year end of the latest available annual report.
G.2.3 Existence of Bonus Plan
This measure comprises a dummy variable, where the existence of a bonus plan, as
ascertained from the annual report, is coded as ‘1’ and non-existence is coded as ‘0’.
G.2.4 Executive Incentive Remuneration
This variable comprises the bonus or incentive specific executive director
remuneration, as disclosed in the latest available annual report, expressed as a
component of total executive director remuneration.
G.2.5 Change in Gross Margin
{ } Explanatory detail overleaf…
Mean EPS Forecast (t) – Actual EPS (t -1)
Actual EPS (t -1)
X 100
Gross Margin (t) – Gross Margin (t -1)
Appendix G
___________________________________________________________________________
130
As the gross margin of a company is already stated in percentage terms, the change in
this variable is calculated by subtracting the gross margin of a company in the
previous period from that of the latest available financial period.
G.2.6 Change in Net Margin
{ } The change in the net margin is calculated, consistent with the gross margin measure.
G.2.7 Change in Gearing
{ } The measure of gearing employed in this study is defined as total debt expressed as a
percentage of total assets, commonly defined as leverage (Thomson One Banker,
2012). This measure of gearing was chosen as alternative measures resulted in
instances of missing data when utilising the Thomson One Banker database. As this
measure is already stated in percentage terms, the change is calculated by subtracting
the level of gearing in the previous period from that of the latest available financial
period.
G.2.8 Change in Average Trade Receivables Collection Period
{ } As the average trade receivables collection period is stated in terms of days, the
change in this variable is calculated in a manner consistent with the abovementioned
variables.
G.2.9 Proportion of INEDs to the Total Board of Directors
This variable is stated in percentage terms, where the total number of independent,
non-executive directors, as ascertained from the annual report, is expressed as a
percentage of the total board of directors.
Net Margin (t) – Net Margin (t -1)
Gearing (t) – Gearing (t -1)
T/Rec Days (t) – T/Rec Days (t -1)
Appendix G
___________________________________________________________________________
131
G.2.10 Proportion of INEDs to the Audit Committee
This variable is stated in percentage terms, where the total number of independent,
non-executive directors on the audit committee, as ascertained from the annual report,
is expressed as a percentage of the total audit committee.
G.2.11 No. of Governance Non Compliance Issues
This variable comprises the number of reported non-compliance issues with the UK
Combined Code on Corporate Governance, as ascertained from the annual report.
G.2.12 No. of Audit Committee Meetings
This variable comprises the number of audit committee meetings held during the
latest financial period, as ascertained from the annual report.
G.2.13 Auditor Type
This measure comprises a dummy variable, where a Big 4 auditor, as ascertained from
the annual report, is coded as ‘1’ and a non-Big 4 auditor is coded as ‘0’. At the time
of this study, a Big 4 auditor includes PricewaterhouseCoopers, KPMG, Deloitte and
Ernst & Young.
G.2.14 Audit Specific Fee(s)
This variable is stated as a ratio, where the total audit specific fee(s) and audit fee(s)
pursuant to legislation, as ascertained from the annual report, are expressed as a
component of total revenue, acting as an appropriate control for size. While previous
studies control for size using the natural logarithm of audit fees (Frankel et al, 2002),
revenue is utilised in this instance, to avoid excessive use of the natural logarithm of
key variables (see Natural Logarithm of Total Assets below).
G.2.15 Beta Value
This measure is obtained directly from the Thomson One Banker database for the date
of 07 June 2012 - providing a measure of the sensitivity of a stock’s price to the
movement of an index (Thomson One Banker, 2012).
G.2.16 Natural Logarithm of Total Assets
Consistent with prior studies, including Frankel et al (2002), this study utilises the
natural logarithm of a key variable to control for size – in this case, total assets.
Appendix G
___________________________________________________________________________
132
G3. Standard Assumptions for the OLS Multiple Regression Model
Where the assumptions underlying multiple regression analysis hold true, a strong
justification for the use of ordinary least squares estimation exists (Newbold, 1988,
p.499). In an ideal scenario, conditions will support all such assumptions; however, it
is unlikely that all assumptions are fully valid in any single regression (Webster,
1995). The primary assumptions underlying such analysis, as outlined by (Newbold,
1988 and Webster, 1995) include:
1. The model is correctly specified, depicting real-world behaviour, generating
results that are not at significant variance with underlying reality.
2. The error observations (𝜀i) are not correlated with one another, i.e. there is no
autocorrelation amongst them.
3. The regression error term (𝜀i) is a random, normally distributed variable with a
mean of zero.
4. All errors have the same variance, i.e. the condition of equal variance in errors
(homoscedasticity) exists.
5. The number of observations (n) exceeds the number of independent variables (k)
by at least two.
6. There is no multicollinearity (significant linear relationship) between the
independent, explanatory variables.
According to Newbold (1988, p.501), where the abovementioned assumptions hold
true, by virtue of the Gauss-Markov theorem, the regression model and least squares
estimators are said to be the best, linear, un-biased estimators.
G.3.1 Assumption One – Model Specification
Where the regression model utilised differs substantially from actual practice, any
conclusions drawn from tests undertaken may be subject to error (Newbold, 1988,
p.566). Specification bias is present where the model, as designed, is inconsistent with
its theoretical foundations, often through the omission or inclusion of certain variables
(Webster, 1995, p.720). Misspecification in previous earnings management models is
well documented, with correlated omitted variables in instances of extreme financial
performance, while attempts at misspecification mitigation procedures have often
reduced test power (Dechow et al, 2011).
Appendix G
___________________________________________________________________________
133
The accrual reversal framework of Dechow et al (2011) provides an apparent solution
for mitigating model misspecification, yet there is no generally identified optimal
model for detecting or examining earnings management activity. The multivariate
regression model employed in this study for examining the determinants of earnings
management is consistent with prior research, where no instances of misspecification
have been identified.
G.3.2 Assumption Two: Error Observations Uncorrelated - No Autocorrelation
Proceeding with an OLS regression model where there are autocorrelated errors (the
error observations are not independent of each other) can have severe implications,
where resultant inferences relating to hypotheses and confidence levels are potentially
misleading (Newbold, 1988, p.593). The Durbin-Watson statistic is utilised to detect
the presence of autocorrelation, with values of this measure close to two generally
indicating that autocorrelation is not a problem (Webster, 1995, p.572).
The Durbin-Watson statistic, calculated for this study on Multiple Regression One
(Section 6.4.4) is measured as follows, where, at the 1% significance level:
DL = 1.539 : DU = 1.813
K = 14 : N = 204
The resultant Durbin-Watson statistic of 1.878 is greater than DU (1.813) above,
therefore the null hypotheses of no autocorrelation amongst the error observations is
accepted (Newbold, 1988, p.587), indicating that autocorrelation is not an issue with
the model employed.
G.3.3 Assumption Three: Normally Distributed Error Term – Zero Mean
O’ Mahony (2010) states that a regression error term with a mean other than zero
results in biased coefficient estimates. The regression error term, calculated on
Multiple Regression One (Section 6.4.4), has a mean of zero (4.2E-16) and a standard
deviation of one (0.999), complying with underlying OLS assumptions. The
assumption of normal distribution is also maintained, indicated by the un-skewed
bell-curve and frequency distribution in Figure G.1 overleaf.
Appendix G
___________________________________________________________________________
134
Figure G.1 – Frequency Distribution and Bell-Curve (Multiple Regression One)
G.3.4 Assumption Four: Homogeneity of Variance
Where the error terms do not all have equal variance, a model is said to exhibit a
degree of heteroscedasticity (Newbold, 1988, p.576). According to Webster (1995,
p.754), where heteroscedasticity exists, the regression coefficients become less
efficient and more unreliable. As the error terms are estimated by the residuals, visual
inspection of a scatter plot where the residuals are plotted against the expected y
values is utilised to determine if any pattern exists amongst the residuals (Newbold,
1988, p.576). As no such discernible pattern exists in Figure G.2 below, the null
hypothesis of homoscedasticity (homogeneity of variance) is accepted.
Figure G.2 – Scatter Plot of Regression Residuals (Multiple Regression One)
-4
-3
-2
-1
0
1
2
3
4
5
-0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3
Reg
ress
ion
Sta
nd
ard
ised
Resid
ual
Regression Standardised Predicted (Y) Value
Scatter Plot of Regression Residuals
Appendix G
___________________________________________________________________________
135
G.3.5 Assumption Five: (N) Exceeds (K) By at Least Two
Webster (1995, p.701) states that in multiple regression analysis, there are (k + 1)
parameters to be estimated and that in order to retain at least one degree of freedom in
the model, N must exceed K by at least two. This requirement is significantly
exceeded in this study, where: N (No. of observations) = 204 and K (No. of
independent Variables) = 14.
G.3.6 Assumption Six: No Multicollinearity between Independent Variables
Multicollinearity arises where there is strong intercorrelation between the independent
variables in the model employed (Newbold, 1988, p.570) and its presence violates one
of the conditions for multiple regression analysis, with resultant inflation of the
standard errors of the coefficients (Webster, 1995, p.717). While there is no
pre-determined cut-off point at which correlation is determined to be excessive in
testing for multicollinearity (Webster, 1995, p.717), generally accepted practice
suggests that a Pearson correlation coefficient of 0.8 between any two independent
variables indicates the presence of multicollinearity. The highest degree of
intercorrelation between the independent variables in this study remains well below
this threshold, measured at 0.2526 between the Change in Gross Margin and the
Change in Net Margin, as indicated in Table G.2 overleaf. Multicollinearity may also
be detected through variance inflation factor analysis, which measures the degree of
multicollinearity contributed by each independent variable, where a VIF result of at
least 10 indicates the presence of multicollinearity (Webster, 1995, p.719). As
indicated in Table G.1 below, the VIF values of all continuous independent variables
are well below this threshold, demonstrating clearly that multicollinearity is not
present in the model employed.
Table G.1 – Variance Inflation Factor Analysis Results
Variable VIF Variable VIF
EPS Growth (%) 1.14 Change in Trade Rec. Days 1.06
Earnings (EPS) Surprise (%) 1.20 INEDs to Total Board (%) 1.18
Exec. Incentive Remuneration 1.05 INEDS to Audit Committee (%) 1.36
Change in Gross Margin (%) 1.21 # Gov. Non-Compliance Issues 1.48
Change in Net Margin (%) 1.24 # Of Audit Committee Meetings 1.06
Change in Gearing (%) 1.08 Audit Specific Fee(s) 1.15
Appendix G
______________________________________________________________________________________________________________________________
136
Table G.2 – Test for Multicollinearity: Continuous Independent Variables - (MAX CORRELATION = 0.2526)
CORRELATION Earnings
Quality
EPS
Growth %
EPS
Surprise
%
Exec.
Incentive
Rem.
Change in
Gross
Margin
Change in
Net
Margin
Change in
Gearing
Change in
T/R Days
INEDs to
Total
Board
INEDs to
Audit
Committee
# Gov. Non
Compliance
Issues
# Audit
Committee
Meetings
Audit
Specific
Fee(s)
Earnings Quality 1.000 -0.094 0.101 -0.048 -0.063 0.090 -0.161 -0.030 0.147 0.045 -0.023 0.031 -0.042
EPS Growth % -0.094 1.000 -0.125 0.016 0.035 0.200 -0.014 -0.188 -0.068 0.035 -0.070 -0.027 0.033
EPS Surprise % 0.101 -0.125 1.000 0.096 -0.231 -0.111 -0.083 -0.019 -0.128 0.013 -0.077 0.038 -0.186
Exec. Incentive Rem. -0.048 0.016 0.096 1.000 -0.058 0.051 0.017 -0.001 0.056 -0.059 0.019 -0.024 0.070
Change in Gross Margin -0.063 0.035 -0.231 -0.058 1.000 0.253 0.051 0.095 0.077 0.015 -0.064 -0.071 0.240
Change in Net Margin 0.090 0.200 -0.111 0.051 0.253 1.000 -0.137 0.029 0.068 0.015 -0.163 -0.034 -0.018
Change in Gearing -0.161 -0.014 -0.083 0.017 0.051 -0.137 1.000 0.024 -0.076 -0.063 -0.047 0.021 0.141
Change in T/R Days -0.030 -0.188 -0.019 -0.001 0.095 0.029 0.024 1.000 -0.021 0.036 -0.022 0.002 0.069
INEDs to Total Board 0.147 -0.068 -0.128 0.056 0.077 0.068 -0.076 -0.021 1.000 0.146 -0.229 0.170 -0.022
INEDs to Audit Committee 0.045 0.035 0.013 -0.059 0.015 0.015 -0.063 0.036 0.146 1.000 -0.487 0.056 -0.104
# Gov. Non Compliance
Issues -0.023 -0.070 -0.077 0.019 -0.064 -0.163 -0.047 -0.022 -0.229 -0.487 1.000 0.012 0.076
# Audit Committee
Meetings 0.031 -0.027 0.038 -0.024 -0.071 -0.034 0.021 0.002 0.170 0.056 0.012 1.000 -0.025
Audit Specific Fee(s) -0.042 0.033 -0.186 0.070 0.240 -0.018 0.141 0.069 -0.022 -0.104 0.076 -0.025 1.000