Studienreihe der Stiftung Kreditwirtschaft an der Universität Hohenheim
Verlag Wissenschaft & Praxis
Matthias Johannsen
Stock Price Reaction to
Earnings Information
Stock Price Reaction to Earnings Information
Studienreihe der Stiftung Kreditwirtschaft an der Universität Hohenheim
Herausgeber:
Prof. Dr. Joh. Heinr. v. Stein
Band 47
Matthias Johannsen
Stock Price Reaction to Earnings Information
Verlag Wissenschaft & Praxis
Bibliografische Information der Deutschen Nationalbibliothek
Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.d-nb.de abrufbar.
ISBN 978-3-89673-481-5
© Verlag Wissenschaft & Praxis Dr. Brauner GmbH 2010 D-75447 Sternenfels, Nußbaumweg 6 Tel. +49 7045 930093 Fax +49 7045 930094
verlagwp@t-online.de www.verlagwp.de
Alle Rechte vorbehalten
Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt.
Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt ins- besondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen.
Printed in Germany
TABLE OF CONTENTS 5
Table of Contents
Table of Contents ... 5
Tables... 8
Figures ... 9
Appendices ... 10
Abbreviations and Variables... 11
1. Problem outline ... 17
2. Investigation of the post-earnings-announcement drift in Germany .... 21
2.1 Introduction... 21
2.2 Current state of academic discussion... 22
2.2.1The concept of market efficiency ... 22
2.2.2The post-earnings-announcement drift ... 22
2.2.2.1Implications for market efficiency ... 22
2.2.2.2Summary of evidence ... 23
2.2.3 Existing explanations for the post-earnings-announcement drift... 26
2.2.4Commentary on existing explanations... 27
2.3 Research design ... 29
2.3.1Research methodology... 29
2.3.2 Considered variables for a risk-based investigation ... 33
2.3.3Hypotheses development ... 36
2.4 Sample construction... 37
2.4.1Data collection ... 37
2.4.2Variable computation... 37
2.4.2.1Unexpected earnings... 37
2.4.2.2Unexpected earnings portfolio formation... 40
2.4.2.3Risk-related variables ... 41
2.4.2.4Abnormal returns ... 43
2.5 Data analysis ... 45
2.5.1Investigation of hypothesis 1 ... 45
2.5.1.1 Cumulative abnormal returns of unexpected earnings portfolios... 45
2.5.1.2 Significance tests of the cumulative abnormal returns... 50
2.5.2Investigation of hypothesis 2 ... 58
TABLE OF CONTENTS
6
2.5.3Investigation of hypothesis 3 ... 60
2.5.3.1Analysis of covariance... 61
2.5.3.2 Sub-samples based on size and book-to-market ratio ... 66
2.6 Conclusion ... 72
3. Investigation of the capital market reaction to earnings management .... 75
3.1 Introduction... 75
3.2 Current state of academic discussion... 76
3.2.1General concept of earnings management... 76
3.2.2Incentives of earnings management... 77
3.2.2.1Types of incentives ... 77
3.2.2.2 Incentives of opportunistic earnings management ... 78
3.2.2.3Incentives of informative earnings management... 81
3.2.3Capital market reaction to earnings management... 82
3.2.3.1Concealing effect of earnings management... 82
3.2.3.2Informative effect of earnings management... 83
3.2.4Summary ... 84
3.3 Research design ... 84
3.3.1Theoretical considerations ... 84
3.3.2Hypotheses development ... 86
3.3.3Research methodology... 88
3.4 Sample construction... 90
3.4.1Data collection ... 90
3.4.2Variable computation... 91
3.4.2.1Earnings management variables ... 91
3.4.2.1.1 Different approaches to measure earnings management ... 91
3.4.2.1.2 Comment on the different approaches ... 95
3.4.2.1.3 Used measures of earnings management ... 98
3.4.2.2Earnings informativeness variables ... 100
3.4.2.3Control variables... 107
3.4.2.4Absolute cumulative abnormal returns... 109
3.5 Data analysis ... 111
3.5.1Correlations among key variables ... 111
3.5.2Investigation of hypotheses 1 and 2... 113
3.5.3Investigation of hypothesis 3 ... 118
3.5.3.1Hypothesis tests ... 118
3.5.3.2Fixed effects panel regressions... 123
3.6 Conclusion ... 134
TABLE OF CONTENTS 7
4. Summary ... 137
4.1 Main results... 137
4.2 Implications for market participants ... 138
4.3 Implications for accounting standard setters ... 139
4.4 Implications for future research... 139
References ... 141
Appendices ... 155
TABLES
8
Tables
Table I: Summary of studies related to the post-earnings-
announcement drift... 25
Table II: T-tests for CAR and CAR differences of the SUE and SRM deciles ... 52
Table III: Estimations for duration and magnitude of the post-earnings- announcement drift... 57
Table IV: Relationship between unexpected earnings and risk-related variables ... 58
Table V: Analysis of covariance of cumulative abnormal returns ... 61
Table VI: Trading strategy returns for size and book-to-market sub-samples ... 67
Table VII: Descriptive statistics and average firm size within sub-samples.... 70
Table VIII: Studies related to earnings management... 80
Table IX: Studies related to pricing of informativeness... 86
Table X: Earnings management incentives variables ... 101
Table XI: Computation of indicator variables... 103
Table XII: Factor loadings obtained from principal component analysis ... 104
Table XIII: Computation of earnings informativeness measure INFO_PCA.. 105
Table XIV: Cluster centroids of k-means cluster analysis... 106
Table XV: Key variables... 111
Table XVI: Correlations among key variables... 112
Table XVII: Investigation of hypotheses 1 and 2... 114
Table XVIII: Investigation of hypothesis 3 ... 119
Table XIX: Fixed effects panel regressions ... 127
FIGURES 9
Figures
Figure I: Timing of the event study ... 30
Figure II: Structure of the analysis of cumulative abnormal returns ... 32
Figure III: Post-earnings-announcement drift ... 46
Figure IV: Post-earnings-announcement drift, anticipation effect ... 48
Figure V: Post-earnings-announcement drift after succeeding announcement... 49
Figure VI: Separation of sample into sub-samples... 89
Figure VII: Separation of sample to investigate hypotheses 1 and 2... 114
Figure VIII: Separation of sample to investigate hypothesis 3 ... 119
APPENDICES
10
Appendices
Appendix I: Distribution of event dates in a random sub-sample... 155
Appendix II: T-tests for CAR of SUE and SRM deciles... 156
Appendix III: Analysis of covariance, long event windows... 159
Appendix IV: Distribution of event dates in random sub-sample ... 160
Appendix V: Random effects panel regressions... 162
ABBREVIATIONS AND VARIABLES 11
Abbreviations and Variables
α = population regression slope coefficient
A = abnormal returns
a = estimated sample regression slope coefficient
abs. = absolute
ACT = total current assets
AktG = Aktiengesetz, German Stock Corporation Act APT = Arbitrage Pricing Theory
) (•
AR = autoregressive regression specification
ARCH Test = Autoregressive Conditional Heteroscedasticity Test
AT = total assets
ATGLNZ = Z-value of natural logarithm of the gross total assets growth rate
avg. = average
β = in regression: population slope coefficient
= in CAPM: market beta
βˆ = estimator for regression slope coefficient b = estimated sample regressions slope coefficient
ME
BE = ratio of the book value of common equity to the market value of common equity
c = regression intercept
CAPM = Capital Asset Pricing Model CAR = cumulative abnormal returns
CDAX = index of all German shares in the Prime Standard and General Standard of the Frankfurt Stock Exchange CEO = chief executive officer
CEQ = cash and cash equivalents
CEQGZ = Z-value of cash and cash equivalents growth rate CFO = cash flow from operating activities
CSE = book value of common shareholders’ equity
CSEGZ = Z-value of book value of common equity growth rate
ABBREVIATIONS AND VARIABLES
12
Δ = difference of first order
δ = decision function in a statistical decision problem
D = binary variable
d = decision in a statistical decision problem DEP = depreciation, amortization and depletion
df = degrees of freedom
diff. = difference
DST = short-term debt and current portion of long-term debt
DT = total debt
ε = regression residual
[ ]
•E = expectation operator
EBIT = earnings before interest and taxes
ed. = editor
EMGMT = earnings management measure DAC
EMGMT_ = earnings management measure based on discretionary accruals EPS = earnings per share
EstG = Einkommensteuergesetz, German Income Tax Law et al. = et alii, et aliae, and others
et seq. = et sequens, and the following page et seqq. = et sequentes, and the following pages
exp = exponentiation with respect to the base e EUR = Euro currency
EXTRZ = extreme financial performance measure F = risk factor in multifactor model of APT
= F-statistic
γˆ = coefficient estimates in Hausman test )
(•
G = function taking on values strictly between zero and one GAAP = generally accepted accounting principles
GoB = Grundsätze ordnungsmäßiger Buchführung, German principles of proper accounting
ABBREVIATIONS AND VARIABLES 13
H = hypothesis
HGB = Handelsgesetzbuch, German Commercial Code
HML = book-to-market variable in Fama/French Three-Factor Model i = cross-sectional identification index
IFRS = International Financial Reporting Standards INC = reported earnings (net income)
INFO = earnings management informativeness measure CL
INFO_ = earnings management informativeness measure based on cluster analysis
PCA
INFO_ = earnings management informativeness measure based on prince- pal component analysis
IPO = initial public offering
j = identification index
k = identification index
) ˆ(•
λ = estimated inverse Mills ratio LCT = total current liabilities
LEVGZ = Z-value of leverage growth rate, where leverage is measured as the debt to assets ratio
LM-Test = Lagrange Multiplier Test
m = binary variable indicating the observability of the dependent variable
ME = market value of common equity
MOMENTUM = returns between 8 and 2 months prior to earnings announcement
υ = integration variable
n = sample size
NDAC = non-discretionary accruals
no. = number
Ω = parameter space of the parameter Q
OI = operating income
OLS = ordinary least squares )
(x
Φ = standard normal cumulative distribution function )
φ(x = the standard normal density function
ABBREVIATIONS AND VARIABLES
14
( )•
P = probability
p = in statistical tests: probability
= in references: page
PAD = post-earnings-announcement drift
PC = principal component from principal component analysis PCA = principal component analysis
par. = paragraph
lim
p = probability limit
PPE = gross property, plant and equipment PRC = provisions for risks and charges
( )•
ψ = reward function in a statistical decision problem Q = parameter in a statistical decision problem q = outcome of the parameter Q
ρ = correlation coefficient
R2 = coefficient of determination Rf = risk-free interest rate
Rm = return on the market portfolio Rp = return on a portfolio of stocks
r = raw stock return
rE = equity cost of capital
RE = residual earnings
REC = net receivables
σ = standard deviation
σˆ = estimator for population standard deviation
s = sample standard deviation
SALES = sales revenues
SDSALESAT = standard deviation of the sales to assets ratio SMB = size variable in Fama/French Three-Factor Model SRM = security return model unexpected earnings SUE = standardized unexpected earnings
ABBREVIATIONS AND VARIABLES 15
τ = range of event window event period
t = time period (year, month, day) in statistical tests: t-statistic T = number of time periods in a panel regression
TAC = total accruals
ü = time de-meaned panel regression residual
UK = United Kingdom
US = United States
V = random variable in a statistical decision problem v = outcome of the random variable V
vol. = volume
w = measurement error
WCP = working capital
x = explanatory variable in regression x = vector of explanatory variables y = in regression: dependent variable
= in figures: time period (year) yˆ = fitted regression values
Z = in statistical tests: Z-statistic in variable computation:
Z-transformation z = variable of interest
PROBLEM OUTLINE 17
1. Problem outline
This work investigates the reaction of capital market participants to information contained in financial statements of public limited companies. Demand for such capital market research originates from several sources. Kothari (2001: 106 et seqq.) names fundamental firm valuation, the role of accounting numbers in con- tracts as well as in the political process and tests of market efficiency as the main sources of demand.
Investors and creditors are interested in the value of a firm to make lending and investment decisions. Their use of information contained in financial statements is an important part of the markets dynamics which drive the price of a security to its fair value, that is, to the expected present value of the future cash flows to which the security entitles its holder. Empirical capital market research in this context helps the investors to spot mispriced securities and make profitable in- vestments. The legal environment in which a firm operates is determined in the political process and shaped by policy makers. In Germany, disclosure standards for example are set by law in the German Commercial Code (Handelsgesetzbuch, HGB, §§ 238 – 315a). The policy makers aim to determine disclosure require- ments in order to satisfy the information interests of the company, its sharehold- ers, its creditors, its employees as well as those of potential investors or the gen- eral public. Empirical capital market research in accounting assists the standard setter in ascertaining whether the stated objectives are met and whether proposed new standards can be expected to serve their purpose. For example, value rele- vance of earnings numbers is seen as an objective of disclosure, and value rele- vance studies are aimed at testing this property of earnings numbers. The ques- tion of market efficiency is of great relevance to all market participants since se- curity prices determine the allocation of wealth between the related parties.1 The present work can be placed in the context of shareholder and investor de- mand for empirical capital market research. It investigates how market partici- pants react to financial statement information. In particular, it examines in two event studies how stock prices are influenced by earnings numbers. Model devel- opment as well as implications for disclosure standards are not topics of the analysis. Verrecchia (2001) reviews the underlying theory and Holthausen/Watts (2001) present an overview of research regarding the value relevance of reported earnings in the context of accounting standard setting.
Looking from the perspective of shareholders and investors, earnings numbers are central for the determination of the fair value of a share. Net earnings (after interest payments) represent that portion of the value generated over the respec-
1 A definition of market efficiency is given in section 2.2.1.
PROBLEM OUTLINE
18
tive period to which shareholders are entitled.2 For this reason earnings numbers attract so much attention in firm valuations. Penman (2004: 121) notes that earn- ings as the result of accrual accounting can in principle capture value added in operations which cash flows cannot. Accruals are non-cash measures of value used in accrual accounting to match the value outflows with the corresponding value inflows. At the same time, due to the discretion required for the estimation of certain accruals, they are often used as instruments for earnings management activities (Kothari 2001: 161). Therefore, the effect of those accruals on the in- formativeness of earnings is open.
This work explores these issues and is structured in two main parts. The first part studies the share price reactions to earnings announcements, and in the second part the share price reaction to earnings management is examined. The literature shows that in opposition to Fama’s (1970, 1991) definition of an efficient market as one in which “security prices reflect all available information” stock price re- actions to earnings announcements take place only gradually. This phenomenon has become known as the post-earnings-announcement drift (PAD). The first part of this work tests for the existence of the drift on the German stock market and scrutinizes the explanatory power of several risk-related variables regarding the drift.
In the second part the effect of earnings management on the informativeness of earnings is examined. As the “purposeful intervention in the external financial reporting process with the intent of obtaining some private gain” (Schipper 1989:
92), earnings management is a relevant factor influencing the informativeness of earnings, which is defined as the degree to which current earnings indicate future earnings. Guay/Kothari/Watts (1996: 83) distinguish between opportunistic earn- ings management and informative earnings management to achieve a more reli- able measure of firm performance. Based on this concept, it is tested whether capital market participants combine earnings management and earnings informa- tiveness information to discriminate between opportunistic and informative earn- ings management.
Following innovations of this work contribute to the research. Testing for the ex- istence of the post-earnings-announcement drift on the German stock market and investigating the effect of risk-related variables regarding the drift produces out- of-sample evidence with respect to studies based on data from the US stock mar- kets. Incorporating a wide range of risk-related variables, which have not been examined simultaneously in previous research, allows of testing their incremental explanatory power. Further, the use of an analysis of covariance in addition to significance tests of a drift-based trading strategy increases the robustness of the findings. Regarding the return reaction to earnings management, the chosen re-
2 Since dividends are paid out immediately, they do not contribute to the value of a share.
PROBLEM OUTLINE 19
search approach is innovative. Two multivariate measures of the informativeness of earnings management are developed. Additionally, a research design is se- lected which allows testing of whether market participants combine earnings management and earnings informativeness information to discriminate between opportunistic and informative earnings management.
The remainder of this work is structured as follows. Chapter 2 investigates the post-earnings-announcement drift and chapter 3 examines the capital market re- action to earnings management. These two chapters are structured in the same pattern. Following an introduction in the respective first section, the current state of the academic discussion is summarized in the second sections. The research design and the development of the hypotheses to be tested are described in the third sections. The following fourth sections are concerned with the data collec- tion and the computation of the employed variables. In the respective fifth sec- tion, the results of the data analysis are discussed and each chapter terminates with a conclusion as the sixth section. A general summary is given in chapter 4.
References and appendices are included in the last two parts of the work.