Equity Valuation and Negative Earnings - The Case of The Dot - Com Bubble (PDFDrive)

Accounting, Finance, Sustainability, Governance & Fraud:
Theory and Application
Ana Paula Matias Gama

Liliane Cristina Segura
Marco Antonio Figueiredo Milani Filho
Equity Valuation
and Negative
Earnings
The Case of the dot.com Bubble
Accounting, Finance, Sustainability,
Governance & Fraud: Theory and Application
Series editor
Kiymet Tunca Caliyurt, Iktisadi ve Idari Bilimler Fakultes, Trakya University
Balkan Yerleskesi, Edirne, Turkey
This series acts as a forum for book publications on current research arising from
debates about key topics that have emerged from global economic crises during the
past several years. The importance of governance and the will to deal with
corruption, fraud, and bad practice, are themes featured in volumes published in the
series. These topics are not only of concern to businesses and their investors, but
also to governments and supranational organizations, such as the United Nations
and the European Union. Accounting, Finance, Sustainability, Governance &
Fraud: Theory and Application takes on a distinctive perspective to explore crucial
issues that currently have little or no coverage. Thus the series integrates both
theoretical developments and practical experiences to feature themes that are
topical, or are deemed to become topical within a short time. The series welcomes
interdisciplinary research covering the topics of accounting, auditing, governance,
and fraud.
More information about this series at http://www.springer.com/series/13615

Ana Paula Matias Gama Liliane Cristina Segura
•
Marco Antonio Figueiredo Milani Filho
Equity Valuation
and Negative Earnings
The Case of the dot.com Bubble
123
Ana Paula Matias Gama Marco Antonio Figueiredo Milani Filho
Management and Economics Department State University of Campinas—UNICAMP
University of Beira Interior Limiera, São Paulo
Covilhã Brazil
Portugal
Liliane Cristina Segura

Universidade Presbiteriana Mackenzie
São Paulo
Brazil
ISSN 2509-7873 ISSN 2509-7881 (electronic)

Accounting, Finance, Sustainability, Governance & Fraud: Theory and Application
ISBN 978-981-10-3007-9 ISBN 978-981-10-3009-3 (eBook)
DOI 10.1007/978-981-10-3009-3
Library of Congress Control Number: 2016956819
© Springer Nature Singapore Pte Ltd. 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #22-06/08 Gateway East, Singapore 189721, Singapore
When you see reference to a new paradigm you should
always, under all circumstances, take cover… There was
never a paradigm so new and as wonderful as the one
that covered John Law and the South Sea Bubble … until
the day of disaster.
(J.K. Galbraith, The Great Crash, 1955)
Foreword
The twenty-first century started with a financial bang, as the bubble built during the
previous decade was severely punctured. Volatility reached new heights and the
dot.com mania fizzled out as the number of IPOs became much scarcer. It is no
surprise that the capital available for the financing of start-up companies suffered a
significant reduction.
However, the world economy carried on, since the emerging economies, espe-
cially the so-called BRIC countries, seemed prone to capture a much larger share
of the world wealth. Financial innovation was not deterred and the risk-sharing
techniques reached a new stage with the widespread dispersion of apparently safe
Collateralized Debt Obligations. The puncture of the real estate bubble triggered by
the subprime crisis which followed led to the most severe downturn since the Great
Depression, with ramifications yet to be tamed. Public policies aimed at controlling
the side effects of the financial crisis changed the rules in ways that had not been
seen before—when had we last witnessed long-lasting negative interest rates out-
side, perhaps, of Japan?
This is a challenging framework to attempt an examination into the valuation of
most assets, let alone start-ups. However, it is more important than ever, as new
clues are definitely needed in such turbulent times. A contrarian investor would not
like to miss opportunities that may be ignored by the larger crowd. An entrepreneur
may similarly sense market opportunities that large corporations may feel too shy to
explore.
This book provides a powerful insight into a very timely issue—how can we
value the most elusive of assets: new ventures at a time of high uncertainty? By
addressing this topic, this work builds upon previous reflections and models from
several leading authors in corporate finance. Damodaran addressed “the dark side of
valuation” to suggest bold procedures to discount future cash flows projected from a
very thin experience, while ignoring the “irrelevant” negative values of the past.
The Gordon formula which permeates the many diverse valuation models requires
serious adaptations to meet the current needs of investors and entrepreneurs alike, if
vii
viii Foreword
these are to be provided with the equity required to carry on with their initiatives, as
debt gets even more out of reach in this volatile environment.
It is no surprise that current investors, be they wealthy individuals or more
seasoned venture capitalists, feel insecure with the basic “comparables” and ratios
often used to support investment and negotiations. Due diligence exercises, how-
ever extensive and resource consuming, often remain sterile without stronger val-
uation tools.
The thorough testing of a large number of companies and the use of models such
as Ohlson’s yield a significant contribution to academics as well as practitioners.
Hard questions, such as the apparently illogical link between negative results and
high capitalization, find a plausible explanation in this study. As this research
shows, rather than reflecting market “irrationality”, such a relation may be due to
the implicit valuation of expenses in research and development or advertising for
new brands that may have been registered as costs and may contribute to the
generation of positive cash flows in the future.
This book sheds significant light on one of the most pressing “unknowns” in
corporate finance, as identified by Braley and Myers: how investment decisions are
carried out in practice. However, the benefits of clearing up this issue go well
beyond the academic interest of corporate finance scholars, to serve economic
agents with key decisions to undertake or even regulators aiming at designing
efficient and reassuring norms that calm markets and channel resources to their most
productive applications.
ISCTE-IUL José Paulo Esperança

May 2016 Dean of ISCTE/IUL Lisbon Business School
Avª das Forças Armadas
Lisbon
Contents
Part I Introduction
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1 Initial Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The Internet: History and Concepts . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 The Internet and Electronic Commerce . . . . . . . . . . . . . . . . . . . . . 5
1.4 e-Business Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Dot.com Companies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 NASDAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.7 Economic Bubbles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.8 The Valuation of Internet Companies . . . . . . . . . . . . . . . . . . . . . . 10
1.9 Objectives and Main Research Contributions . . . . . . . . . . . . . . . . 13
1.10 Organization of the Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Part II Literature Review

2 The Ohlson and Feltham Ohlson Models . . . . . . . . . ............. 19
2.1 The Ohlson Model (OM) . . . . . . . . . . . . . . . . . . ............. 19
2.2 The Extent of the Ohlson Model: The Feltham
Ohlson Model (FOM) . . . . . . . . . . . . . . . . . . . . . ............. 24
2.3 The Effect of Conservatism Accounting . . . . . . . ............. 31
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. 40
3 The Value Relevance of the Variables Earnings and Book
Value of Equity for Valuation Purposes . . . . . . . . . . . . . . . . . . . .... 43
3.1 The Informational Content of the Variable Results . . . . . . . . .... 44
3.2 The Informational Content of the Variable Book Value
of Equity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 52
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 61
ix
x Contents
4 The Impact of Investment in Intangible Assets

on the Market Value . . . . . . . . . . . . . . . . . . . . . . . . . . ............. 65
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. 65
4.2 The Impact of Investment in Intangible Assets
on the Market Value . . . . . . . . . . . . . . . . . . . . . . ............. 66
4.3 The Company Value and the Potential Growth. . ............. 71
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. 80
Part III Empirical Study

5 Period, Sample Selection and Definition of Variables . . . . . . . . . . . . . 85
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.2 Definition of the “New Economy Period—NEP” . . . . . . . . . . . . . 86
5.3 Definition of “Net Firms” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.4 Criteria for Selection of Samples . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.4.1 New Economy Companies: “Net Firms” . . . . . . . . . . . . . . 87
5.4.2 Companies with a IPO Date Contemporaneous
to “Net Firms”: “Non-Net Firms” . . . . . . . . . . . . . . . .... 88
5.5 Composition and Comparative Analysis of the Two
Samples: “Net Firms” and “Non-Net Firms” . . . . . . . . . . . . .... 91
5.5.1 The Samples of “Net Firms” and “Non-Net Firms” . .... 91
5.5.2 Comparative Analysis of “Net Firms”
Versus “Non-Net Firms” . . . . . . . . . . . . . . . . . . . . . . .... 91
5.6 Definition of Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 96
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 105
6 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
6.2 The Effect of Life Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
6.3 Criteria for Subdivision of Samples . . . . . . . . . . . . . . . . . . . . . . . 116
6.4 Methodology of Fama and Macbeth . . . . . . . . . . . . . . . . . . . . . . . 124
6.5 Research Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
7 Analysis and Discussion of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.1 Discussion and Analysis of Results . . . . . . . . . . . . . . . . . . . . . . . 139
7.1.1 Econometric Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.1.2 Validation of Empirical Research Hypotheses . . . . . . . . . . 140
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Part IV Conclusions
8 Conclusions and Suggestions for Further Research . . . . . . . . . . . . . . 161
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
List of Figures
Figure 2.1 The relationship between the multiples P/B and PER
and future abnormal earnings (FE) and current
earnings (CE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 35
Figure 3.1 The relationship between the value of the company’s
equity and results due to the liquidation option held
by shareholders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 48
Figure 3.2 a Relationship between the variables profits and returns
according the principle of prudence. b Relationship
between the variables profits and returns given
the liquidation option by shareholders . . . . . . . . . . . . . . . . . . .. 50
Figure 4.1 The effect of investment in R&D and advertising
in abnormal future profitability based on the networking
effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 76
Figure 5.1 Identification of outliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 90
Figure 6.1 Trend analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Figure 6.2 Dow Jones Internet Composite Index—Monthly Closing
Prices in the period July 1997 to February 2002 . . . . . . . . . . . . 122
Figure 7.1 a, b Evolution of the number of firms per sample
depending on the reported value of the variable results . . . . . . . 142
Figure 7.2 Relationship between the estimated coefficients after
controlling for BVE variable (model 6.3) and without
the control for BVE variable (model 6.5) . . . . . . . . . . . . . . . . . 148
xi
List of Tables
Table 3.1 Summary of the factors supporting the weak correlation

between the variable results (positive) and the movement
of prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 47
Table 3.2 McCallig (2004) criteria for splitting the sample . . . . . . . . . . .. 57
Table 3.3 Summary of the various studies on the information
content of the variables earnings and book value
of equity for evaluation purposes . . . . . . . . . . . . . . . . . . . . . . .. 59
Table 4.1 Relationship between MVE and the investment
in R&D and advertising . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 70
Table 5.1 New economy firms listed on ISDEX . . . . . . . . . . . . . . . . . . . .. 88
Table 5.2 Sample sizes of net firms: systematization
of other researches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 89
Table 5.3 Composition of the sample of net firms . . . . . . . . . . . . . . . . . .. 92
Table 5.4 Composition of the sample of non-net firms
(see footnote of Table 5.3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Table 5.5 The number of IPOs by market and by year . . . . . . . . . . . . . . . . 94
Table 5.6 Composition of net firms and non-net firms by industry . . . . . . . 95
Table 5.7 Number of companies listed by market. . . . . . . . . . . . . . . . . . . . 96
Table 5.8 Variables Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Table 6.1 Trend analysis for the sample of net firms . . . . . . . . . . . . . . . . . 111
Table 6.2 Trend analysis for the sample of non-net firms . . . . . . . . . . . . . . 114
Table 6.3 Differences between means/medians in the sample
of net firms: companies with profits and companies
with losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Table 6.4 Differences between means/medians in the sample
of non-net firms: companies with profits and companies
with losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Table 6.5 Split of the samples according to business model . . . . . . . . . . . . 124
Table 7.1 Annual regressions for the sample of net firms (model 6.5) . . . . 141
Table 7.2 Annual regressions for the sample of non-net firms
(model 6.5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
xiii
xiv List of Tables
Table 7.3 Correlation matrix—sample of net firms . . . . . . . . . . . . . . . . . . . 144

Table 7.4 Correlation matrix—sample of non-net firms . . . . . . . . . . . . . . . 145
Table 7.5 Annual regressions for the sample of net firms
(model 6.3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
(model 6.3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
(model 6.6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
(model 6.6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
(model 6.7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
(model 6.7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
List of Appendices
Appendix 2.1 Deduction of the OM Model (Eqs. 2.8 and 2.9) . . . ....... 36
Appendix 2.2 Deduction of the Equivalence of the Preposition
Number 1 of Feltham and Ohlson (FOM) (1995)
Model and Eqs. 2.15a, 2.15b and 2.15c . . . . . . . . . ....... 38
Appendix 2.3 Deduction of the Feltham and Ohlson
(1995)—Eq. 2.22 . . . . . . . . . . . . . . . . . . . . . . . . . . ....... 39
Appendix 5.1 Number of net firms by SIC . . . . . . . . . . . . . . . . . . ....... 99
Appendix 5.2 Number of Non-Net Firms by SIC . . . . . . . . . . . . . . . . . . . . 101
Appendix 6.1 The CUSUM Test for the Variable
“Results Res_IExt”—Net Firms . . . . . . . . . . . . . . . . . . . . . . 132
“MVE/BVE”—Net Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
“Results Res_IExt”—Non-net Firms . . . . . . . . . . . . . . . . . . 134
“MVE/BVE”—Net Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Appendix 6.5 Percentage of Firms with R&D Expenses
Greater than Sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
xv
Part I
Introduction
Chapter 1
Introduction
Abstract The Internet has changed our lives enormously. In the late 1990s, the
Internet constituted a technological revolution and created huge expectations for the
new business models. Many entrepreneurs, supported by venture capital, have made
fortunes, based on potential profits from their high-tech start-ups. Despite the fact
that traditional models of economic evaluation did not present evidence justifying
the great appreciation that the newly established companies were generating, many
investors were nevertheless betting on future earnings to offset present losses. In
this chapter, we provide an overview about the speculative bubble involving
internet companies (dot.com or net firms) during 1999–2003, and we present the
objectives and main research contributions of this study. Additionally, we present
the structure of this book.

Keywords Internet Dot.com companies Valuation of internet firms Economic
bubbles
1.1 Initial Comments
The new economy sector was an integral part of the vertiginous growth of stock
market prices by the end of the 1990s. Fama and French (2002) registered an annual
rate of return of 8.81% in Standards & Poor’s 500 from 1972 to 2000. From 1995 to
1999, the Center for Research in Security Prices (CRPS) registered an annual
growth rate of 24%. Smithers and Wright (2000) reported Tobin’s Q ratios for
capitalization of the US stock market which had never before been observed in the
twentieth century.
Shiller (2000) calculated the highest level ever for the multiple net profit (price
earnings ratio—PER), adjusted based on the ten-year moving averages, superior
even to the relative values of the period 1901–1929.
The elevated market capitalizations were influenced by social media which
created a euphoric atmosphere associated with the period’s innovation (Shiller
2000). Demers and Lewellen (2003) drew attention to the impact of the marketing
© Springer Nature Singapore Pte Ltd. 2017 3

A.P. Matias Gama et al., Equity Valuation and Negative Earnings, Accounting,
Finance, Sustainability, Governance & Fraud: Theory and Application,
DOI 10.1007/978-981-10-3009-3_1
4 1 Introduction
campaigns associated with the Initial Public Offering (IPO) process, especially for
the net firms.
Although the phenomenon of a positive valuation of losses is significant in the
universe of Internet companies, it is not entirely new. Amir and Lev (1996) iden-
tified a similar behaviour in the mobile phone sector in the 1980s. Between 1984
and 1993, the 14 American mobile operators recorded negative results in 69% of
the trimesters. Moreover, this percentage was even higher in the biotechnology
sector (72%).
In view of these facts, analysts, professionals and academics, as well as the press,
questioned the suitability of traditional evaluation methods, especially in the con-
text of net firms during the period of the “dot.com bubble”. So, a multiplicity of
ratios based on web-variable traffic proliferated, e.g. “time spent browsing on the
given site”, “number of sites visited,” “percentage of internet users”, on the grounds
that, since they were better able to capture the value chain of the net firms, these
variables measured the network effect boosted by the World Wide Web space (www)
more easily.
It was assumed, especially, that the increase of the Internet users would increase
the turnover and reduce the unit fixed cost, making it easier to forecast sales and
profitability of an emerging sector which showed a high level of losses.
Several authors have challenged this perspective, emphasizing that the source of
value could be summarized as “growth” and “profitability”.
Since the collapse of the dot.com bubble in 2001, the situation for start-ups has
changed considerably, especially in relation to funding streams (which shrank
significantly), as well as the expectations of economic agents regarding the risks of
investing in this type of business.
In general, this significant market reversal generated a resource-shortage cycle
and risk aversion that resulted in the creation of more appropriate methodologies to
deal with technology-based start-ups. These not only minimized risks, but also
minimized the financial losses associated with attempts to create innovative
solutions.
The general idea is that start-ups are likely to fail and, to minimize the costs
(psychological and financial) of these failures, entrepreneurs must deal with the
start-up as a high-risk experiment and base their actions on processes that seek to
minimize the waste of resources and reduction of learning time.
1.2 The Internet: History and Concepts
Wars of great scale, such as World War II, as well as others of smaller scope,
epitomized the risk inherent in centralized computer networks. If the server was
bombed, all stations would lose connection.
In the 1960s, during the so-called Cold War, ARPA (Advanced Research
Projects Agency) belonging to the US Department of Defense, initiated the
development of ARPANET, a network characterized by decentralization, so that if
1.2 The Internet: History and Concepts 5
one of the knots was hit and suffered damage, communication between other
computers was still possible using other surviving connections.
In 1982, the transmission protocols and TCP/IP (Transmission Control Protocol
and Internet Protocol) address became operational and the network was renamed
the Internet (Intercontinental Network).
Gradually, the network gained greater penetration in academia, linking univer-
sities and enabling a better exchange of information and experiences among
researchers.
Since 1989, with the fall of the Berlin Wall, the priorities that fostered the
Arpanet project have been evolving; in 1992, the US government ended state
control of the Internet and allowed it to be developed with funds from the private
sector.
The World Wide Web (www), designed at the European Particle Physics Lab as a
vehicle for sharing information on high-energy physics between physicists working
on a dispersed international environment, adopted a standard called Hypertext
Markup Language (HTML).
The Web resulted in a much more intuitive and user-friendly Internet, enabling
its use by non-technical and non-academic people.
In 1990, the ARPANET ceased to exist, and commercial-service providers
began to emerge, i.e. companies connected to the Internet and which enable sub-
scribers to link their computers for network access.
From the onset of the World Wide Web, the Internet has experienced an evo-
lution never seen before by other means of communication. In the US, for example,
the Internet reached 50 million users within only four years, whereas to achieve this
number of users it took the personal computer 16 years, television 13 years and the
radio 38 years. In 2000, about 400 million people worldwide had access to the
Internet.
With the subsequent development of new technologies such as mobile tele-
phones connected to the World Wide Web at lower costs, the number of electronic
users expanded in an unprecedented way.
1.3 The Internet and Electronic Commerce
Due to new technologies, business have adapted to available tools, targeting

improvement of their business models and their relationship with various stake-
holders. The use of Internet-based platforms for commercial operations was a
natural consequence of the opportunities resulting from this new context.
Given that supply and sales transactions were the first enterprise tasks to be
carried out, the term “e-commerce” became the popular term for such a virtual
environment. However, the potential for electronic tasks is not restricted to pur-
chase and sales transactions, so a broader concept, e-business, became recognized
and understood.
6 1 Introduction
Thus, the scope of the term e-business is greater than that of e-commerce, since
the former includes not only the e-commerce, but also the contact and support
activities which form the main mechanism of business. E-commerce is not just
about commerce transactions or buying and sales over the Internet. It is a global
reset strategy of the old business models with the aid of technology, adding value to
the company as well as to its related parts.
One can consider, therefore, that e-business favours processes for direct contact
between customers and suppliers, as well as it enables market analysis, investment
analysis, the search of information about the macro-environment, market research,
etc. The use of the Internet and other networks and information technologies
function in support of e-commerce, communications and cooperation between
companies, and business performance on the web, plus in support of internal
company communication, as well as connections with clients and business partners.
As for e-commerce, we can define it as the buying and selling of information,
products and services through computer networks.
1.4 e-Business Environment
Internet-technology-oriented business (e-business) improves productivity and

increases the efficiency of various companies, delivering easier and faster com-
munication with partners, suppliers and customers. e-business has been character-
ized as a complex fusion of business processes, enterprise applications, and the
organizational structure needed to create a high-performance business model.
The electronic business environment enables different types of relationship, as
shown below
• G2G: Government-to-Government—applications that enable integration
between public agencies.
• B2G: Business-to-Government—applications that enable integration between
business and government and vice versa.
• G2C: Government-to-Consumer—applications that enable integration between
the government and consumers in general.
• B2C: Business-to-Consumer—applications that enable integration between
business and private consumers.
In the beginning, the companies sought to put down their banners on the Internet
only to signal their presence, not knowing how to produce a return. Later, emphasis
was placed on the order flow and revenue and the relationship with different
stakeholders.
The electronic business applications are part of the information management
context, serving both the internal applications of the organization, and those
applications linking partners, suppliers, and virtual communities. Some denomi-
nations are
1.4 e-Business Environment 7
Business Intelligence (BI): is the production of management information to support

decisions about search, analysis and application of qualitative and quantitative
information activities, which results in knowledge about the logic of the target
market and about taking actions that are strategic in nature.
Knowledge Management (KM): focused on intellectual-capital management and
knowledge of the organization. In short, it is the process for obtaining, managing
and sharing the experience and expertise of employees. The goal is to have access
to better information at the right time, using technology on a corporate level and
scope.
Supply Chain Management (SCM): is the integration of the various companies that
comprise the supply chain, in all its stages. Thus, the SCM does not consider only
the physical production process, but the flow of information and capital between all
companies involved in the chain. Hence, the communication constitutes a key
element for the maintenance and supply chain management.
Customer Relationship Management (CRM): aimed at the management of all
contacts made by the company with its customers. Generally, its main goal is to
place the customer at the centre of business processes in order to create that kind of
perception that enables us to anticipate the current and potential needs of that
customer. The goal is to understand and strengthen relationships with the con-
sumers of its products.
Efficient Consumer Response (ECR): corresponds to the initiatives to improve the
efficiency of the entire chain. In this sense, industrial and commercial enterprises, as
well as other supply chain members (logistics operators, banks, equipment and
vehicle manufacturers, computer companies, etc.) work together in pursuit of
meeting common standards and efficient processes that minimize costs and optimize
productivity in their relationships.
Collaborative Planning, Forecasting, and Replenishment (CPFR): is a program that
aims to be a valid alternative to the Efficient Consumer Response (ECR) taking
advantage of its positive aspects, but presenting solutions towards the demands.
ECR has a greater focus on the supplier, especially in reducing costs and stream-
lining businesses, so, the objective focus of CPFR is on the final consumer by
means of joint management of processes and information exchange.
1.5 Dot.com Companies
Organizations that offer their products and do their business predominately on the
Internet are Dot.com companies. They are so named because it is a simple and
direct way to refer to the commercial domain (.com) at the end of their Web
addresses.
In the late 1990s, dot.com companies based their business model on the number
of accesses, visibility or even the creation of virtual communities. Therefore,
intensive disclosure was essential for the image, and the expenditure of the
8 1 Introduction
abundant venture capital available at that time provided the means to support this
activity. The goal was to generate a competitive advantage aiming at the generation
of compensatory profits in the future. TV ads in prime time, luxurious offices and
acquisition of traditional companies were common examples of exorbitant spend-
ing. However, few companies survived to be able to tell the story.
Razi et al. (2004) studied the main features of many dot.com companies that
went bankrupt during the speculative bubble and also identified the characteristics
of those who successfully overcame this moment to survive. The potential causes
for failure were separated into two main categories: controllable and uncontrollable
causes.
Controllable causes are those under direct control of decision-makers and were
divided into strategic, operational and technical causes. On the other hand, the
uncontrollable causes are beyond the influence of the managers and were divided
into technical and behavioural causes.
Among the most significant companies related to bubble, we highlight:
• Boo.com
• Books-a-Million
• Broadcast.com
• e.Digital Corporation
• Freeinternet.com
• GeoCities
• TheGlobe.com
• GovWorks.com
• Inktomi
• InfoSpace
• Kozmo.com
• Lastminute.com
• The Learning Company
• Lycos
• MicroStrategy
• Open.com
• Pets.com
• Pixelon
• Startups.com
• Think Tools AG
• Tiscali
• VA Linux
• Webvan
• WorldCom
• Xcelera.com
1.6 NASDAQ 9
1.6 NASDAQ
The NASDAQ (National Association of Securities Dealers Automated Quotation)

is one of the major stock markets of the world and was the first electronic stock
exchange, connecting buyers and sellers directly, as well as providing enabling
trading of the assets of high-tech companies, also called “New Economy”
companies.
It has its origin in a petition of the Congress of the United States to the Securities
and Exchange Commission (SEC), in view of a report on the safety of the markets
which found unregulated markets to be opaque. The SEC proposed its automation
resulting in the creation of NASDAQ. Its first session was on 8 February 1971.
The IPO process on NASDAQ is much simpler and less expensive than on the New
York Stock Exchange (NYSE), and that is why many medium and small businesses
choose to launch their IPOs on NASDAQ.
After a deep restructuring in 2000, NASDAQ became a for-profit company fully
governed by shareholders, issuing and exchanging its own shares in its own market.
It currently has about 2,800 listed companies, and its daily average turnover, in
2015, was approximately US $95 billion.
The NASDAQ Composite Index is the primary stock market index of the US
electronic stock market. It reflects the market value of all shares traded on
NASDAQ. This index is widely used by the financial market as an indicator of the
performance of companies in the technology sector. Since there are US and non-US
companies listed on NASDAQ, the NASDAQ Composite index cannot be con-
sidered an American stocks index. The success of the NASDAQ Composite Index
stimulated the creation of other indexes used as performance indicators of various
segments of the broad market of NASDAQ shares.
Until March 2000, when the NASDAQ index exceeded the level of 5,000 points,
a pervasive demand for risky investments in technology companies was clearly
evident, especially for Internet companies.
1.7 Economic Bubbles
An economic or speculative bubble could be described as a situation in which asset

prices reflect unrealistic expectations about the future. It is “trade in high volumes at
prices that are considerably at variance with intrinsic values” (King et al. 1993).
Considering that it is a very difficult task to identify intrinsic values in the daily
life of the financial and capital markets, bubbles are not generally recognized until
they are bursting, i.e. when asset prices start plummeting.
According to Singh (2013), individual agent behaviour can be rational, but the
framework within which agents operate can produce prices which do not corre-
spond to their fundamental values. Certainly, it is possible that individuals be
strongly influenced by others in their decision-making, and this is generally known
10 1 Introduction
as herd behaviour. Banerjee (1992) defines herding as “everyone doing what

everyone else is doing, even when their private information suggests doing
something quite different”.
Many investors follow the thinking or directions of those around them or the
advice of financial analysts. When specific stocks become popular for implausible
prospects and investors seek to buy large holdings of these stocks, this collective
action creates temporary price pressures which can drive prices up to high levels.
After the large buying momentum ceases and facing the lack of economic reasons
for maintaining high prices, the stock prices then may fall abruptly.
1.8 The Valuation of Internet Companies
During the period of full expansion of the speculative dot.com bubble, Damodaran
(2000) published an article with a quite significant title: The dark side of valuation:
firms with no earnings, no history and no comparables. Can Amazon.com be
valued?
In that article, the author analyzed the obstacles which are faced when assessing
young or embryonic companies: negative or abnormally low earnings, short his-
torical data, and the lack, or small number of, comparable companies. In other
words, limited history, small revenues in conjunction with big operating losses, and
a propensity for failure make the dot.com companies tough to value because the
types of problems appear conjugated.
The calculation of risk parameters, such as the beta ratio (in Capital Asset
Pricing Model—CAPM), is impaired to the extent that one needs to work with a
five-year-minimum time series.
When we have a very short period of time for analysis, it is difficult to capture
variable fluctuations that occur on an annual or longer period and that can be very
significant sometimes.
Another difficulty that arises from the lack of a data set is the inability to test the
reasonability of the evaluation results, i.e. to compare the results of evaluations over
different periods and at earlier times.
The absence of comparable firms is the third complicating factor of dot.com
assessments. By “comparable firms” is meant firms that are really in the same
business and that have similar characteristics and fundamentals. In addition to the
use of prior-period data, analysts often use information about comparable firms in
the evaluations. As a basis for projections, it is common to use risk indexes,
investment volumes, or firms of similar size that work in the same industry.
Damodaran has developed alternative ways and solutions to manage this set of
barriers to the assessments and better to choose among different alternatives. He
begins by addressing each of the problems, and then concentrates on the specific
event in which the conjugated problems arise in the same company.
There are three basic options for approaching the cases of companies with
negative current results, namely: standardization of results; adjustments in projected
1.8 The Valuation of Internet Companies 11
revenues in conjunction with adjustments in operating margins; and reduction of

leverage.
To normalize the results mean to disregard the negative results, considering them
as something abnormal, and to take the average of positive results, return on
investment, return on equity or the profit margin from prior periods (if positive), or
to use those same parameters of comparable firms, the year in which the company
under evaluation had losses. By adopting this procedure, one also adopts the
implicit assumption that the firm had not experienced abnormal behaviour during
the period (or periods) observed.
One has to question the author about this procedure which seems somewhat
arbitrary, based on the “feeling” of the analyst and, for this reason, to question its
technical consistency. The adjustment in revenues and estimated margins is based
on current revenues (which are never negative) and, on the same estimates, one
estimates an operating margin and net margin overtime.
An alternative to this procedure is to estimate the investments in future years and
the return on capital overtime to project the outcome. The estimate-operating
margin is the sustainable margin, i.e. based on the history of the firm and the
average margin of comparable firms. In addition, a period of adjustment in the
operating margin is estimated, which will depend on, among other factors, the
magnitude of the difference between current income and sustainable margins as
well as the reasons for this difference, for example, economies of scale and
investment in infrastructure, with a greater gestation period.
It is common to find operationally viable firms with high levels of debt. The debt
might be due to, among other reasons, significant investments in infrastructure,
generally associated to periods of long maturation, or leveraged buyouts.
The procedure of adjustment of the firm’s leverage consists, initially, in esti-
mating the optimal level of company debt, by a traditional analysis of cost of capital
or, alternatively, to take an average for the specific segment.
There are some ways to adjust the degree of leverage, among which, the fol-
lowing three can be highlighted:
– Postpone investments and redeem the debt with cash generated from
depreciation;
– Let the increase in revenues and operating income push up the value of the firm
(if the debt grows at a slower pace than the firm value growth, then the debt ratio
will fall);
– Raise funds by issuing shares or debentures (or admitting new members with
capital) to pay the debt.
The choice between the different approaches to address the issue of corporate
evaluations with losses involves understanding the reasons for which the results are
negative, the firm’s characteristics (e.g. if there is, or if there is not, a change in its
scale over time); after this, and the procedures outlined above should be applied, as
suitable.
12 1 Introduction
It must be ascertained whether the phenomenon is transient, abnormal, or if the

firm’s business is cyclical; if the losses result from long-term operational problems,
specific to the firm (and not in the sector in which it operates); or even if the poor
results are due to structural problems.
It is not possible to change the fact of the youth of a company or that there are no
comparable companies. So, in the evaluation of companies, the absence or scarcity
of historical data about the evaluated firm is usually compensated by the use of
comparable data and business information; and the absence or insignificance of
comparable companies’ data is balanced with by elements of the available time
series.
There are important considerations to be made on the issue of comparability
between companies. First, one must examine very carefully the issue of similarity of
business, i.e. if the companies really operate with the same product and the same
markets.
Another relevant issue is the richness of information available for each firm and
the stability of these companies. For example, while there is less information about
auto companies than Internet companies, the available information on the former
are much more profound than about the latter.
Finally, the degree to which comparable companies are at different stages of their
life cycles deserves a lot of attention from the analyst. Obtaining data from com-
parable companies in different stages of their life cycles would be ideal.
If the majority of comparable companies are in one stage, for example those
considered “high growth”, there is little information to estimate the rate and time of
deceleration in the transition phase to stability. If all these problems are present at
the same time, in the same firm, this is certainly a great challenge for evaluation,
However, Damodaran was not the first to address the issue of the evaluation of
start-ups.
Ever since the pioneering works of Ball and Brown (1968) and Beaver (1968), the
literature that examines the price/results (positive) is extensive. However, research on
price/losses is scarce, and the empirical results obtained so far are contradictory.
Nevertheless, and in spite of the few studies on this topic, the research carried out
proposes three potential explanations for this phenomenon/anomaly.
Hayn (1995) proposed the first explanation. The losses assume a transitional
nature, if they persist, shareholders exercise the liquidation option they hold on the
company’s assets. Thus, in the context of consecutive losses that indicate a higher
likelihood of financial bankruptcy, the variable “equity” is more relevant for the
assessment, acting as a proxy of the liquidation value of the company.
The second explanation states that the phenomenon of negative relations of
price‐earnings results from an incorrect appraisal of the model for specification—
capitalizing of the results (earnings model) (Collin et al. 1999).
The third explanation holds that the losses by technology companies in start-up
phase are due to high investments in intangible assets, particularly in research and
development (R&D) and advertising (Chan et al. 2001). According to the Generally
1.8 The Valuation of Internet Companies 13
Accepted Accounting Principles (GAAP), investments in intangible assets are

treated at full cost in the year in which they are made.
Several studies have analyzed companies with losses, including those outside the
Internet sector (Hand 2001, 2003; Kozberg 2009; Tokic 2004, 2005). These losses
are often motivated by high investments in intangible assets (Kozberg 2009).
1.9 Objectives and Main Research Contributions
This study is inspired by the valuation models of Ohlson (OM) (1995) and Feltham
and Ohlson (FOM) (1995): where an omitted variable from the model (book value
of equity) is positively correlated to the dependent variable (the market value of the
companys) and negative (positive) to the independent variable included in the
model (net income); the exclusion of this type of variable induces a negative bias
(positive) of the estimated coefficient of the independent variable; here resides the
explanation for the anomaly of “positive evaluation of losses”.
According to the evaluation models Ohlson (OM) (1995) and Feltham and
Ohlson (FOM) (1995), the phenomenon of “positive evaluation of losses” can be
explained by the effect of “conservatism accounting” which leads to the underesti-
mation of assets resulting from not capitalizing on investments in intangible assets
(e.g. R&D and advertising).
Hence, taking as a theoretical framework the evaluation models of OM and
FOM, and given the magnitude of the phenomenon of “positive evaluation of
losses” recorded in the universe of net firms, the central objective of this research is
to analyze the relationship between the stock market capitalization and net income
(loss) reported by the net firms throughout the period of the new economy (NEP). In
particular, this study intends to analyze how the market evaluates the companies
that have registered losses, depending on their magnitude and persistence. The
effect of the life cycle will be examined. In brief, this study seeks to:
(i) To evaluate the effect of “accounting conservatism” on the relationship
between the market value of equity (MVE) and the information reported in
the financial statements of net firms;
(ii) To evaluate, in the context of systematic losses, the relevance of the main
determinants of value (value drivers) to the market value of the equity of
these companies;
(iii) To analyze the suitability of these variables as proxies for growth opportu-
nities, given that investments in R&D and advertising are aimed at creating a
critical customer mass to monetize the network effect generated by the
Internet, creating a brand image (see the examples of Amazon and Yahoo).
The latter to be accomplished by developing software and new platforms to
improve web-site design, e-mail alerts and creating greater security mecha-
nisms for online transactions.
14 1 Introduction
The option for utilizing the evaluation models of OM and FOM is justified by
the fact that these models represent a significant theoretical contribution to the field
of models of business valuation, which is a highly relevant theme in the field of
corporate finance. Based on the pioneering work of Miller and Modigliani (1961),
these authors typified the impact of growth opportunities for assessing corporate
value. But they innovate by introducing: (i) the effect of dynamic information at the
level of abnormal returns and (ii) the impact of non-financial information, according
to the theory of efficient markets.
So, given the dynamic information, which are defined as an autoregressive
process of first order, in the medium term, and given the competition effect, the
abnormal returns tend to converge to the industry average. The effect of
non-financial variables is relevant to the extent that they are evidence of the limi-
tations of financial statements, i.e. their inability to report in a timely manner all
relevant information affecting investor expectations (lack of timeliness). This
information is immediately incorporated into prices, but only later reflected in the
financial statements. In this context, the financial statements underestimate the
present value of growth opportunities held by the company.
The impact of these variables is particularly relevant in emerging sectors, such as
the case of the Internet sector. On the one hand, it is a sector with some techno-
logical complexity; on the other hand, there is very limited historical information,
which enhances the information asymmetry between managers (insiders) and
investors (outsiders).
Considering specifically the analyzed phenomenon, the “positive evaluation of
losses” in the universe of net firms, these models establish that the losses may occur
early in the life cycle of companies, as originally envisaged by Myers (1977).
The justification is based on the fact that only a part of the investment is
capitalized, mainly in intangible assets, the remainder being considered as costs for
the year (conservatism accounting effect), in obedience to GAAP.
However, the market identifies this type of investment according to the proba-
bility of the existence of the future of growth opportunities, which supports the high
expectations of supernormal returns associated with these companies (positive
signals).
Models of OM and FOM demonstrate that growth may increase the expected
value and business results. But due to the conservatism accounting effect, the
company’s value increases more rapidly than the expected value for the results, and
this explains the two indicators: both the multiple results (PER) and the multiple of
book value (M/B market-to-book) tend to record high levels, so the goodwill, as
measured by the difference between the market value and the accounting value of
equity tends to persist even in the medium term.
In fact, the results obtained allow us to conclude that investors
(i) Do not focus their attention for the purposes of evaluation, only in the
variable “results”, as an aggregate variable;
(ii) Value positively the variables R&D and advertising, which for accounting
purposes are treated as costs, associating to these investments the probability
1.9 Objectives and Main Research Contributions 15
of the existence in the portfolio of greater opportunities for growth, so higher

expectations of supernormal returns;
(iii) View the “equity—BVE” variable as particularly relevant for evaluation,
recording the company losses. This is because the market sees the BVE
variable as proxy, as foreseen by the models of OM and FOM, for normal
future results;
(iv) Consider the BVE variable to be itself, also, a tool to reduce agency costs,
particularly with creditors; and identifies it as a proxy for the “recognized
assets”, given the predominance of intangible assets in technology-based
companies;
(v) See increased investment under “R&D” and advertising and in the line of
results of Hayn (1995) and Joos and Plesko (2004), as linked to a change of
the company’s profile in the decade of the 1990s, i.e. small companies,
mostly technology-based, which register losses that tend to assume greater
magnitude and persistence.
In view of these results and the main contributions of this research, we emphasize
that the registration of loss may not be indicative of a process of value destruction. In
clear opposition to the theory of abandonment option, in certain contexts, the
information content of losses is not irrelevant to the assessment; in particular if the
losses are associated to the process of creation/growth performance, i.e. high growth
opportunities. As a result, it is incorrect to evaluate the reported losses of all com-
panies in the same manner, as this may lead to erroneous empirical findings.
Finally, companies in financial stress, particularly technology-based companies,
tend to opt for Mergers & Acquisitions processes (M&A) as a restructuring strategy
to the exemption of bankruptcy, which would imply a greater destruction of value.
1.10 Organization of the Work
We have structured the present book therefore as follows.

Chapter 2 focuses on the analysis of the evaluation models of Ohlson
(OM) (1995) and Feltham and Ohlson (FOM) (1995), which originates from the
model of Gordon, while still incorporating the principles of Modigliani and Miller
(1958) and Miller and Modigliani (1961).
Chapter 3 examines the phenomenon of “positive valuation losses” and the
information-content relevance of variable results and the content of equity. Chapter
4 analyzes the impact of the investments in R&D on the market value of the
company, and evaluates the effect of growth on the market value of the company’s
equity, with a particular focus on the Internet sector. Chapter 5 presents the main
aspects of the methodology defined for the empirical study. Chapter 6 presents the
method. Chapter 7 analyzes the phenomenon of “positive valuation of losses” in
companies of the new economy in the US, and the life-cycle effect of those com-
panies. Finally, in Part III, the main findings of this research as well as for future
research studies are presented.
16 1 Introduction
References
Amir L, Lev B (1996) Value-relevance of nonfinancial information: the wireless communications

industry. J Acc Econ 22(1–3):3–30
Ball R, Brown P (1968) An empirical evaluation of accounting income numbers. J Acc Res 6:159–
178
Banerjee AV (1992) A simple model of herd behavior. Q J Econ 108(3):797
Beaver WH (1968) Market prices, financial ratios and the prediction of failure. J Acc Res 6:179–
192
Chan L, Lakonishok J, Sougiannis T (2001) The stock market valuation of research and
development expenditures. J Financ 56(6):2431–2456
Collin D, Pincus M, Xie H (1999) Equity valuation and negative earnings: the role of book value
of equity. Acc Rev 74(1):29–61
Damodaran A (2000) The dark side of valuation: firms with no earnings, no history and no
comparables. Can Amazon.com be valued? New York: Stern School of Business. http://www.
stern.nyu.edu/*adamodar
Demers E, Lewellen K (2003) The marketing role of IPO: evidence from internet stocks. J Financ
Econ 68(3):413–437
Fama E, French K (2002) The equity premium. J Financ 57:637–659
Feltham G, Ohlson J (1995) Valuation and Clean surplus accounting for operating and financial
activities. Contemp Acc Res 11(2):689–731
Hand JRM (2001) The role of book income, web traffic, and supply and demand in the pricing of
US internet stocks. Eur Financ Rev 5(3):295–317
Hand JRM (2003) Profit, losses and non-linear pricing of internet stocks. In Hand JRM, Lev B
(ed) Intangibles: an Oxford management reader. Oxford University Press, Oxford
Hayn C (1995) The information content of losses. J Acc Econ 20(2):125–153
Joos P-R, Plesko GA (2004) “Valuing loss firms”, MIT Sloan working paper 4491-04. http://ssrn.
com/abstract=562043
King RR, Smith VL, Williams AW, Van Boening M (1993) The robustness of bubbles and crashes
in experimental stock markets. In: Day RH, Chen P (eds) Nonlinear dynamics and evolutionary
economics. Oxford University Press, New York, pp 183–200
Kozberg AR (2009) Explaining revenues and valuations for internet firms: the usefulness of
accounting and non-financial information in explaining revenues and valuations for internet
firms. VDM
Miller M, Modigliani F (1961) Dividend policy, growth and the valuation of shares. J Bus 34
(4):411–433
Modigliani F, Miller M (1958) The Cost of capital, corporate finance and the theory of investment.
Am Econ Rev 48(3):261–297
Myers S (1977) Determinants of corporate borrowing. J Financ Econ 5:147–175
Ohlson J (1995) Earnings, book values, and dividends in equity valuation. Contemp Acc Rev 11
(2):661–687
Razi MA, Tarn M, Siddiqui FA (2004) Exploring the failure and success of DotComs. Inf Manag
Comput Secur 12(3):228–244
Shiller RJ (2000) Irrational exuberance. Princeton University Press, Princeton
Singh V (2013) Did institutions herd during the internet bubble? Rev Q Financ Acc 41:513–534.
Doi:10.1007/s11156-012-0320-1
Smithers A, Wright S (2000) Valuing wall street: protecting wealth in turbulent markets.
McGraw-Hill, New York
Tokic D (2004) R&D, advertising and the market value of internet firms: part I. J Internet Commer
3(2):79–99. Doi:10.1300/J179v03n02_05
Tokic D (2005) R&D, advertising and the market value of internet firms: part II. J Internet
Commer 4(4):23–40. Doi:10.1300/J179v04n04_02
Part II
Literature Review
Chapter 2
The Ohlson and Feltham Ohlson Models
Abstract This chapter analyses the phenomenon of “positive valuation of losses” in

the new economy companies in the US. One of the potential explanations of this
phenomenon is that these companies are start-up companies, mostly technology-
based, that invest massively in intangible assets, in particular research and devel-
opment (R&D) and advertising. Under Generally Accepted Accounting Principles
(GAAP), these investments should be considered at full cost in the year they occur.
Thus, in this chapter, we analyse the Ohlson (OM) (Contemp Acc Rev 11(2):661–
687, 1995) and Feltham and Ohlson (FOM) (Contemp Acc Rev 11(2):689–731,
1995) valuation models. Feltham and Ohlson (Contemp Acc Rev 11(2):689–731,
1995) demonstrated analytically, using dynamic information, that losses, particu-
larly at the stage of start-up in growth and technology-based companies, are con-
sidered to be costs that create an effect of conservatism accounting, consequently,
there is an undervaluation of assets, hence the results and equity. However, this
situation tends to be reversed over time, because given the principle of rationality, the
investors continue to invest in the company if those investments are associated with
abnormal profitability expectations.
Keywords Technology-based companies Positive valuation of losses Ohlson

and Feltham and Ohlson models
2.1 The Ohlson Model (OM)
Beaver (2002: 457) states: ‘The F–O approach [Ohlson 1995 (OM) and Feltham
and Ohlson 1995 (FOM)] is, in my opinion, one of the most important research
developments in the past ten years’. The advantage of the Ohlson model (OM) is
that it defined a conceptual framework that relates the market value of the company

DOI 10.1007/978-981-10-3009-3_2
20 2 The Ohlson and Feltham Ohlson Models
(MVE) with the past and the future financial information of the company, i.e. with
current and future expected net income, with the book value of equity (BVE), and
with dividends.1
The initial theoretical framework of OM was the neoclassical model of dividends
developed by Williams (1938), but known as the Gordon and Shapiro (1956)
model. Gordon postulates that: (i) the growth rate for dividends is constant; (ii) the
preferences/beliefs of the agents are homogeneous; and (iii) they are neutral to risk.
Hence, the dividend model is defined as follows:
X
1
Pt ¼ Rs
f E t ð dt þ s Þ ð2:1Þ
s¼1
where
Pt share price at time t;
dt net dividends paid at time t. The dt variable reflects all net
transactions with shareholders, such as the payment of dividends,
new shares issued to finance new investments, and/or repurchase of
shares. For simplicity, we designate this variable only for dividends;
R = (1 + rf), where rf is the free-risk rate;
Et [.] the expected value operator, conditional on information available at
time t.
In this context, and assuming two principles:
(i) The principle of clean surplus relation (CSR), which states that
bvt ¼ bvt1 þ xt dt ð2:2Þ
where
bvt value of equity at the end of the period t. By analogy, bvt−1 corresponds to the
value of equity in the previous period (t − 1);
xt the net results for the year t.
The variables “bvt” and “xt” are exogenous to the model.2
1
According to Lo and Lys (2001), in 1999, and with reference to the OM model, the mean number
of citations was already higher than nine.
2
Holthausan and Watts (2001) criticize the OM model because it is a partial equilibrium model,
where the financial variables used are defined exogenously to the model. But as Beaver (2002:
458) claims, parsimony is also a very important quality in any model, arguing that: “By analogy,
the capital asset pricing model (CAPM) has the demand for financial institutions, financial
institutions yet we observe empirically”. With this reasoning, Barth et al. (2001: 90) state: To our
knowledge, there is no academic theory of accounting that derives the demand for accounting
2.1 The Ohlson Model (OM) 21
According to the CSR principle, any changes to the book value of the company
(bvt) are the result of income generated and retained in the company, i.e.
Δbvt = xt − dt, where dt reflects all transactions directly with shareholders (i.e.
distribution of dividends, issue of new shares to finance new investment projects
and/or repurchases). The intuition behind this principle is that all transactions
affecting the assets and liabilities of the company, and consequently, the value of
equity, should be reflected in the income statement, and its effect is reflected in the
net income variable. This property, and according to Zhang (2000), reconciles any
changes in the value of the assets held by the company with the flow of income
generated by them. Thus, Ohlson (1995) does not “force” that new investments are
financed only via retained earnings, unlike the closed models of self-sustained
growth (e.g. Gordon’s model). Thus and in accordance with Miller and Modigliani
(1961) the financing of new investment by retained earnings or issuing shares are
perfect substitutes.
Dividends affect the level of capital (bv) in t, but the net income remains
unchanged (xt).3 Analytically:
@bvt = @dt ¼ 1 ð2:3aÞ
@xt = @dt ¼ 0 ð2:3bÞ
@xt = @dt is not obtained directly from (Eq. 2.3a), but is consistent with the same
because @ bvt1 ¼ @ bvt þ @dt @xt ¼ 1 þ 1 0 ¼ 0 (Ohlson 1995: 667).
@dt @dt @dt @dt
Introducing the abnormal variable, defined as:
xat ¼ xt ðRf 1Þbvt1 ð2:4Þ
where xat measures the excess returns that the company receives.
Because the results exceed the cost of capital, Ohlson (1995) expressed the
present value of expected dividends (PVED), based on the net income and the value
of equity. Hence, the PVED is4:
X
1
Pt ¼ Rs
f Et ðxt þ s bvt þ s þ Rf bvt1 þ s Þ
a
ð2:5Þ
s¼1
(Footnote 2 continued)
information arising from the equilibrium forces and provides the mapping of accounting
information into price shares.
3
This assumption is in line with the principle of perfect capital markets, so it excludes any signal
effect associated with the variable “income”.
4
Expressing dividends according to the current results, through the CSR principle and replacing xt,
the expression obtained for the abnormal results, we obtain dt ¼ xat bvt þ Rf bvt1 .
This expression, after some algebraic transformations, allows us to redefine the

PVED model as follows5:
X
1
Pt ¼ bvt þ Rs
f Et ðxt þ s Þ
a
ð2:6Þ
s¼1
This model is well known and reported in the literature as the residual income
valuation model (RIV) or Edwards and Bell (1961) model (White et al. 1997:
1062).
As shown by Lo and Lys (2001), the Gordon’s model and RIV are analytically
equivalent, so to reject the RIV means ignoring that financial assets are a function
of the present value of expected future cash flows.
The innovation of Ohlson (1995) against the RIV model or the Gordon model
lies in the treatment that gives the structure of the time series of abnormal results
ðxat Þ. To define the stochastic process that follows the variable xat , Ohlson (1995)
introduces the variable vt—other information, i.e. a variable that captures relevant
events in terms of information that affects prices, but are not yet reflected in the
financial statements. This time lag of the occurrence of certain events that are
relevant to the formulation of economic agents’ beliefs on the growth of abnormal
results of the company, is one of the limitations ascribed to the financial statements,
or rather to its ability to disclose all relevant information timely, i.e. lack of time-
liness (Rayn 1995; Beaver 2002). To fill this gap, Ohlson (1995) supports his model
in a dynamic information, which he defines as an autoregressive process of first
order, and which features the dynamics of abnormal results. Analytically, the
dynamic information is defined as:

xatþ 1 ¼ wxat þ vt þ e1;t þ 1
ð2:7Þ
vt þ 1 ¼ cvt þ e2;t þ 1
where the parameters w and c are fixed and known and assume values between
]0, 1[.6 In a broad sense, these exogenous parameters to the model are determined
by the environment that characterizes the company. The random terms e1s ; e2s have
Et ðek;t þ s Þ ¼ 0 with k = 1.2 and s 1. The model imposes the independence of vt in
relation to xat because Et ½vt þ s depends only on vt, with vt reflecting all (not just
financial) information relevant to the estimation of abnormal returns, regardless of
their past values. However, its effect is reflected in xat , which is incorporated in the
5
Note that Rs
f Et ðbvt þ s Þ ! 0 com s ! 1, i.e. the present value of capital converges to zero as
the time horizon tends to infinity. The model assumes that the equity grows at a rate less than rf.
6
The parameters w and c assume values greater than zero for economic conditions and values
less
than 1 in order to ensure stability/stationarity of the model. This condition implies that the
Et xatþ s ! 0 and Et ðvt þ s Þ ! 0 with s ! 1. Indeed if w = 1, this means that growth
opportunities persist indefinitely, which is not consistent with the empirical evidence.
2.1 The Ohlson Model (OM) 23
variable bvt, by the property of CSR.7 This dynamic information enables the
company to earn abnormal returns over a period of time, and this effect is captured
by the parameter w. However, due to the competition effect the abnormal process,
i.e. firms’ profitability trend, tends to converge to the average of the economy—
mean reverting.
Combining Eqs. 2.3a and 2.3b, the Eqs. 2.1 (PVED), 2.2 (Principle CSR) and
2.7 (dynamic information), Ohlson (1995: 669) defines the evaluation function
based on the calculation of the expected value for the abnormal results. Hence, the
enterprise value is defined as8:
Pt ¼ bvt þ a1 xat þ a2 vt ð2:8Þ
where
w
a1 ¼ 0 and
Rf w
Rf
a2 ¼ [ 0:
ðRf wÞðRf cÞ
According to the OM, the market value of the company (MVE or Pt) is a linear
function of the level of capital invested in the company (bvt), the abnormal results
ðxat Þ generated by the company and the variable (vt)—other than the financial
information. Ohlson (2000) suggests the analysts’ forecasts for future one-year
results could be a good proxy for the variable vt. The market value of the company’s
equity is much more sensitive to the variables xat and vt, the greater the persistence
of the parameters w and c, exogenous to the model, and therefore α1(w) and α2ðcÞ
are increasing in their determinants.
Equation 2.8 can be reformulated in terms of net profit (adjusted for dividends)
and the value of equity (bvt), where φ corresponds to the multiple of results—price
earnings ratio (PER):
Pt ¼ j½uxt dt þ ð1 jÞbvt þ a2 vt ð2:9Þ
with
(
u ¼ RfR1
f
j ¼ ðRf 1Þa1 ¼ ðRRff1 Þw

w :
7
The CSR defines bvt ¼ bvt1 þ xt dt being xat ¼ xt ðRf 1Þbvt1 ; replacing xt in the CSR
expression we obtain bvt ¼ xat þ Rf bvt1 d t , an expression that shows that any relevant events
from the point of view of information are contained in the value of equity (bvt) through the
“dynamic information”.
8
The analytical deduction of the Eqs. 2.8 and 2.9 appear in the Appendix 2.1.
The previous expression can be interpreted as a weighted average of the eval-

uation model based on the updated income flows—the discounted profits (earnings
model) and the assessment model from stock of assets required for generated the
income flows—the model on equity (book value model).9
Ohlson (1995) also shows that in the medium- and long-term, the variable (bvt)
is an unbiased estimator of the market value of the company (MVE), i.e. there is an
unbiased accounting property. In the short term, Ohlson (1995) admits the exis-
tence of goodwill, which defines as the flow of abnormal results that the company
expects to receive, and which are derived from trademarks, patents, location, cus-
tomer loyalty, investment in R&D, advertising and specificity of the organizational
model—intangible assets, which are potential sources of value creation.
Analytically:
X
1 a
Pt bvt ¼ Rs
f xt þ s : ð2:10Þ
s¼1
Indeed, using the evaluation function defined (Eq. 2.8), it is shown that:

Et ½Pt þ s bvt þ s ¼ a1 Et xatþ s þ a2 Et ðvt þ s Þ ! 0 com s ! 1: ð2:11Þ
i.e. in the medium and long term goodwill is null; in this context the book value of
equity (bvt) is an unbiased estimator of MVE.
2.2 The Extent of the Ohlson Model: The Feltham Ohlson

Model (FOM)
In the work of Feltham and Ohlson (1995) (FOM), the authors introduce two new
effects: (i) the understatement of operating assets, accounting conservatism; and
(ii) growth in the operational assets. The effect conservatism accounting reflects the
persistence of the difference between the market value of equity (MVE) and book
value (BVE), which is the source of the unrecorded goodwill. This unrecorded
goodwill may result due to an understatement of existing and/or an overestimation
of expected abnormal results.
To demonstrate these two effects, the authors continue to assume the neoclas-
sical model of discounted dividends (PVED) and in accordance with Miller and
Modigliani (1961), Modigliani and Miller (1958) the irrelevance of dividend policy
9
In theory, Ferreira and Sarmento (2004) argue that the equity valuation and evaluation on the
basis of updated income streams should give the same value. However, empirically, and given the
existence of goodwill associated with the presence of intangible assets and the relevance or lack of
relevance of financial statements, which derives from their (in)capacity in terms of timely reporting
of all relevant and reliable information, the two approaches tend to have marked differences.
2.2 The Extent of the Ohlson Model: The Feltham Ohlson Model (FOM) 25
and the separation between operating and non-operating activities. The separation
of such activities will have different effects on the evaluation function.
The non-operating activities include assets and liabilities traded in perfectly
individualized markets, whereby the value of this class of asset tends to match its
market value, generating investment with zero net present value (NPV). Therefore,
the evaluation of such assets does not imply any specification, contrary to operating
assets. The difficulties in the assessment of the value of operating assets are related
in that they are not evaluated in a perfect, liquid market.
Assuming the separation between operating and non-operating activities and
including the dividend policy, Feltham and Ohlson (1995) began by defining a set
of accounting and financial variables, from which the evaluation function is
specified.
Thus, considering a multiperiod context, where in each period [t = 0, 1, 2 …],
the company discloses all the information about its operating and non-operating
activities; this information is described by Zhang (2000: 128) as follows:
These data are random prior to their disclosure and the probabilistic structure governing
their stochastic behaviour is exogenous.
In this context the variables considered are:

bvt value of the company’s equity at date t;
xt net profit generated in the period [t − 1, t];
dt dividends at date t;
fat net non-operating assets (non-operating assets minus liabilities) at time t10;
it result from non-operating activities in the period [t − 1, t];
oat net operating assets, i.e. operating assets minus operating liabilities, on the
date t;
oxt operating result for the period [t − 1, t];
ct operating cash flow, i.e. the cash flow generated by operating activities net of
investments;
Pt market value of the company (MVE) on date t.
The relations established between these variables are:
(i) in line with the OM, the principle CSR:
(a) The book value of equity results from the aggregation of operating (oat)
and non-operating assets (fat). Analytically: bvt = oat + fat;
(b) Similarly, the results generated by the two types of activities (operational
—oxt, and not operational activities—it), i.e. xt = oxt + it;
(c) Consistent with the OM model, the principle that defines CSR:
10
This variable can take a negative value when the non-operating liabilities exceed the
non-operating assets. For convenience of analysis, and similarly to the dt variable, it is considered
fat > 0.
bvt ¼ bvt1 þ xt dt ;
(ii) Net interest relation (NIR)—Income generated from non-operating activities:
it ¼ ðRf 1Þ fat1 where Rf ¼ 1 þ rf : ð2:12Þ
The rate considered is the free-risk rate (rf), which is independent of the
financial situation of the company (i.e. fat is >0 or fat <0).11 As it is assumed
that the non-operating assets and liabilities are paid at the free-risk rate, this
type of activity generates a net present value (NPV) of zero. The basic
intuition of this reasoning is based on the principle that non-operating assets
and liabilities are traded in perfect markets, which resemble a cash account
(numeraire asset) measured without any risk (Morgenstern 1963).
(iii) Financial asset relation (FAR)—Relationship between net non-operating
assets:
fat ¼ fat1 þ it ½dt ct : ð2:13Þ
At the beginning of the period, the company begins its activity with a volume
of non-operating assets—fat−1. During the period t, these assets generate an
income it, as dividends paid only at the end of the year t. The cash flow
generated by operating activities (cash flow to the firm—ct) is also deter-
mined at the end of the period. Note that the amount [ct − dt] affects the level
of non-operating assets at the end of the year, but not the level of income
generated in the period (it). The variable ct gains particular relevance in the
FOM model, compared to the OM model.
(iv) Operating asset relation (OAR)—Relationship between net operating assets:
oat ¼ oat1 þ oxt ct : ð2:14Þ
The reasoning behind this relationship is similar to the CSR principle. The
company starts its activity with a certain level of operational assets (oat−1),
which generate an outcome in the period—operational results (oxt); the cash
flow (ct) generated by operating activities is transferred to non-operating
activities.12 Given the FAR relation, the transference of ct for financial
11
In a context of perfect markets, the company cannot change interest rates. Moreover, given the
homemade concept, individual investors cannot mimic the decisions of indebtedness of the
company, since, at a higher or lower level of debt, the company is not a creative source of value
Miller and Modigliani (1961), Modigliani and Miller (1958). To emphasize that, Feltham and
Ohlson in 1999 incorporated risk aversion and the existence of heterogeneous preferences of
investors in their article.
12
If negative, ct corresponds to net investments in operating assets (oat).
activities does not generate any gain or loss, because such transference is
taken at market value, thus the ct variable is objectively measured indepen-
dently of any accounting principles underlying the valuation of operating
assets (oat) (Feltham and Ohlson1995).
Since the criterion for decision-making by managers should be creating wealth,
the aim of Feltham and Ohlson (1995) is to determine the value of the company and
not the value that is distributed to shareholders (e.g. dividend model). Thus, with
reference to the Gordon’s model (PVED), in which the value is determined by the
level of expected wealth transfer to shareholders ðdt þ s Þ, and considering the rela-
tionships between various financial accounting-defined variables (i.e. CSR, NIR,
FAR and OAR), Feltham and Ohlson (1995) demonstrate the equivalence of the
neoclassical model (PVED) with the following three expressions (Proposition 1)13:
X
1
Pt ¼ fat þ Rs
f Et ðct þ s Þ; ð2:15aÞ
s¼1
X
1 a
Pt ¼ bvt þ Rs
f E t xt þ s ; ð2:15bÞ
s¼1
X
1 a
Pt ¼ bvt þ Rs
f Et oxt þ s ; ð2:15cÞ
s¼1
Equation 2.15a, and according to Miller and Modigliani (1961), shows that the
key factor to generating value for a company is the net present value of cash flow
flows from operating activities and the risk that is inherent them. Equation 2.15b is
equivalent to the PVED model (Eq. 2.6). Equation 2.15c specifies that, given the
principle of separation between operating and non-operating activities, abnormal
results derive from operating activities. Only the investments in real assets that have
the capacity to create value, given the existence of economic agents in the markets
with expertise in specific business and with inside information and skills that enable
them to track innovation, thus sustaining their competitive advantages in the face of
competition.
Therefore, in the terminology of Feltham and Ohlson (1995), and in line with
Miller and Modigliani (1961), Modigliani and Miller (1958), the value of the com-
pany is MVEt = fat +(oat + gt). From this equation, it follows that goodwill is a
function only of the expected results from abnormal operational results, i.e. it depends
only on the operational activities.
13
The analytical deduction of these formulas and their equivalence with the PVED model are
reported in Appendix 2.2.
The demonstration is sustained by the CSR principle, and the NIR and FAR
relations. Therefore:
Pt ¼ gt þ fat þ oat ,
ð2:16Þ
Pt fat ¼ oat þ gt
using Eq. 2.15a, we obtain
X
1
aot þ gt ¼ Rs
f Et ðct þ s Þ ð2:17Þ
s¼1
By Eq. 2.15c:
X
1 a
aot þ gt ¼ oat þ Rs
f Et oxt þ s ,
s¼1
ð2:18Þ
X
1
gt ¼ Rs
f Et oxatþ s
s¼1
This expression has two implications: (i) goodwill (gt) can be different from zero,
i.e. the value of the operating assets (oat) differs from its current expected cash
flows; and (ii) the bias from the difference between market value (MVE) and book
value (BVE) does not differ just in the short term, as it tends to persist in the
medium and long term. It is from the persistence of this differential that the problem
of unbiased accounting versus conservative accounting arises.14
In this context, Feltham and Ohlson (1995) define the property unbiased
accounting as occurring when the Et ðgt þ s Þ ! 0 com s ! 1. As a negation of this
property, there is conservative accounting, whereby the Et ðgt þ s Þ 0 com s ! 1.
Thus the property unbiased accounting (Proposition 2) implies:
" #
X
1
Et ðoat þ T Þ ¼ Et Rs
f Et þ T ðct þ T þ s Þ ! 0 com T ! 1;
s¼1
14
For Ohlson (1995) unbiased accounting is characterized by a self-corrective process, in which
the medium- and long-term abnormal results tend to zero and the return on equity ratio
(ROE) converges to r (the cost of capital). Kothari (2001) highlights this autocorrective effect of
the OM model, which makes the OM “immune” to the manipulation of policies and/or accounting
principles. For example, any strategy to increase the results in t, will increase the bvt. In the
following period, this effect tends to be offset by a reduction of the abnormal results because the
cost of capital [(Rf − 1) * bvt−1] increases.
or equivalently,
" #
X
1
Et Rs
f Et þ T ðoxt þ T þ s ÞÞ
a
! 0 com T ! 1: ð2:19Þ
s¼1
Replacing (→0) by (>0) occurs the conservative accounting effect. In other

words, unbiased accounting occurs if the average value of operating assets (oat)
equals the present value of future cash flows, or the abnormal operating results
value is zero with τ → ∞. Hence, the proposition 2 suggests that conservatism
accounting understates the book value of operating assets and/or overstates the
present value of the abnormal of expected operating results.15
With the aim of considering the effect of the persistence of abnormal results, the
effect of conservatism accounting, the growth in operating assets (oat), and oper-
ational results (oxt), Feltham and Ohlson (1995) redefine the dynamics of infor-
mation initially specified by the OM model. Thus, the dynamic information (linear
information model—LIM) becomes defined as:
8 a
>
> ox ¼ w11 oxat þ w12 aot þ v1t þ e1;t þ 1
< tþ1
oat þ 1 ¼ þ w22 oat þ v2t þ e2;t þ 1
ð2:20Þ
>
> v ¼ c1 v1t þ e3;t þ 1
: 1;t þ 1
v2;t þ 1 ¼ c2 v2t þ e4;t þ 1

where Et ej;t þ s ¼ 0 with j = 1, 2, 3 and 4 and τ > 0.
The dynamic of information depends now not only on the persistence of
abnormal results, the effect captured by the parameter w11, as well as the growth of
investments in operating assets (reflected in the parameter w22), but also on the
effect of conservatism accounting at the level of operational assets via parameter
w12. The v1t and v2t variables, and in line with the OM model, aim to capture all
available information that may change investors’ expectations about the persistence
of abnormal and growth resulting from new investments. Thus, the vector (v—other
information) can be interpreted, assuming the semi-strong principle of efficiency,
i.e. the prices tend instantaneously to reflect all publicly available information /m
t
(Fama 1976). If we subdivide /m into two, i.e. economic and financial infor-
t
mation ( F/m t and other information NF/m t ), then the vector v is identified with
the latter subset of information, i.e. other non-financial information not yet incor-
porated in the financial statements. However, the FOM model emphasizes, that the
P1 s P1
s¼1 Rf Et ðct þ s Þ ¼ oat þ
15
This effect is most easily seen in the expression: s¼1
Rs a
f Et oxt þ s deduced from the Eq. (2.15c). Given the objectivity in how the variable ct (cash
flow to the firm) is measured (which is independent of any criteria or accounting policy), an
understatement of operating assets (oat) must be compensated by an overestimation of abnormal
results ðoxat Þ; thus the value of the variable ct remains unchanged.
effect of these parameters is limited to the short term. In this context, the dynamic
evolution of cash flows is characterized by:

ct þ 1 ¼ w11 oxat þ Rf w22 þ w12 oat þ ½v1t v2t þ e1;t þ 1 e2;t þ 1 ð2:21Þ
The dynamics of cash flows depend on the current abnormal results of the w11,
w12 parameters that measure the understatement of operating assets (conservatism
accounting), the increase in operating assets (w22), and even the information that is
being made available to the market (vector—v).
Restrictions that can be imposed on the parameters of the LIM are:
(i) jch j 1 with h = 1.2.
The purpose of this restriction is to ensure that random events, whose effect
is captured by v1t and v2t variables, and not reflected in the financial state-
ments,
have no impact on the medium and long term, i.e.
Et vh;t þ s ! 0; com h ¼ 1; 2 e s ! 1:
(ii) w11 , the parameter that measures the persistence of abnormal results assumes
values ε ]0, 1[. With this condition, the objective is to introduce restrictions
to the persistence of abnormal results. Thus, w11 0 eliminates potential
oscillations that would not be economically acceptable and w11 1, permits
abnormal results that persist for some time, but its effect declines over time
because the competition within the medium term tends to eliminate abnormal
returns.
(iii) The parameter w22 , which reflects the effect of the increase in operating
assets assumes values within the range ]1, Rf[ with Rf = (1 + r). This
restriction imposes limitations on the growth of operating assets in the long
term to ensure convergence
when calculating the present value of operating
abnormal results oxat and the expected cash flows (ct).
(iv) The parameter w12 enables us to introduce the dichotomy unbiased
accounting ðw12 ¼ 0Þ versus conservatism accounting ðw12 0Þ, i.e. the
problem of underestimation of operating assets.16
Given the reformulation of the LIM of the FOM, and continuing to take the
neoclassical framework of PVED and relationships previously established between
the variables (i.e. CSR, NIR, FAR and OAR), the evaluation function is now
defined as:
16
To impose w12 0 the model eliminates the effect of aggressive accounting, i.e. the MVE is
less than BVE. This restriction simplifies the analysis, and is in accordance with the empirical
evidence. Indeed, Feltham and Ohlson (1995: 701) find, and taking as reference the data from
the Compustat database, that the MVE of companies tends to be greater than 2/3 of their BVE.
Stober (1999) found the opposite (aggressive accounting), but only for the period 1973–1979,
and justified as a consequence of the oil shock.
Pt ¼ bvt þ a1 oxat þ a2 oat þ b vt ð2:22Þ
with
w11
a1 ¼
Rf w11
w12 Rf
a2 ¼
ðRf w22 ÞðRf w11 Þ

Rf a2
b ¼ ðb1 ; b2 Þ ¼ ; :
ðRf w11 ÞðRf c1 Þ Rf c2
In this context, goodwill (gt) corresponds to:
Pt bvt ¼ gt ¼ a1 oxat þ a2 oat þ b vt ð2:23Þ
that is goodwill is an increasing function of operating abnormal results, whose

persistence is measured by the parameter w11 (the higher w11, the higher is α1),
operating assets (oat) only when they are undervalued because the necessary con-
dition for α2 > 0 is that w12 > 0 and, the variable vt ðvÞ. We should also notice that
in both models (OM and FOM) the tax effect is ignored. However, given that both
models assume a perfect capital market (i.e. the models do not admit the existence
of costs derived from asymmetric information, agency costs and transaction costs),
then the tax effect will not have a material effect on the evaluation function.17
2.3 The Effect of Conservatism Accounting
According to Richardson and Tinaikar (2004), the accounting conservatism effect

can occur in a company as a result of investments in items such as R&D and
advertising. In accordance with GAAP, this type of investment should be recorded
as historical costs and recognized immediately as expenses (ex ante conservatism)
which underestimate the present value of cash flows from these investments.
Given the company profile examined in this research namely, new economy
companies18 and companies with an Initial Public Offering (IPO) date that is
contemporary with net firms—non-net firms (see Sect. 5.4.2), also which are firms
that operate mainly in high-tech sectors (high-tech firms), we will focus our analysis
on the first kind of conservatism accounting (ex ante) identified by Richardson and
Tinaikar (2004)—the impact of immediate accounting as costs of investment in
17
The deduction of this model (Eq. 2.22) is in Appendix 2.3.
18
The concept of the new economy company is defined in Sect. 5.4.1.
intangibles such as R&D and advertising on the variables “MVE”, “BVE” and “net
income”, the main determinants of value according to the OM and FOM.
Feltham and Ohlson (1995) show that growth companies, particularly technol-
ogy companies in the early stages of their life cycle (start-up companies), where
investments in intangible assets (e.g. R&D and advertising) predominate, have
generally poor results or even negative results, since only a portion of their
investments are capitalized, while the investments such as R&D and advertising are
immediately recognized as an expense in the income statement. Consequently, and
because these investments tend to persist over time, in addition to the understate-
ment of the variable “results”, the variables “assets” (unrecognized assets) and
“equity” (BVE) are undervalued (McCrae and Nilsson 2001).
So consider the time t = 0, where the initial investment made by shareholders is
(−d0), and is applied in non-operating assets (fa0), so −d0 = fa0 and bv0 => 0.
Then, if P0 = do at this time, the goodwill (g0) is null. The company continues to
invest in operating assets (oat), and, by the OAR, the variable cash flow is negative,
i.e. c1 < 0, as consequences of these investments. Considering the Proposition 2
(Eq. 2.19), which defines the property unbiased accounting as:
" #
X
1
Et ðoat þ T Þ ¼ Et Rs
f Et þ T ðct þ T þ s Þ with T ! 1;
s¼1
or equivalently,
" #
X
1
Et Rs
f Et þ T ðoxt þ T þ s Þ
a
! 0 with T ! 1;
s¼1
P1 s
it is possible for a period of time T ε [0, ∞[, that the s¼1 Rf E0 ðoxs Þ\0,
a
implying that the company can expect to get negative results in the first years of its
life, as the result of pursuing high-growth opportunities, particularly if they are
associated with investments in R&D and advertising (counted as costs). However,
this situation tends to be reversed, because in the future the company will only
continue to undertake new investment projects if their expectations are associated
with abnormal returns; otherwise we are looking at the effect of free cash flow
(Jensen 1986). However, this phenomenon is more typical in companies in the
maturity phase.
To better clarify the analysis, the authors define the Proposition 9, which uses on
date 0 a null goodwill:
P0 ¼ d0 ¼ fa0 with oa0 ¼ ox0 ¼ 0 but the E0 ðoa1 Þ [ 0: ð2:24Þ
Thus, assuming conservative accounting, i.e. the understatement of operating

assets:
2.3 The Effect of Conservatism Accounting 33
E0 ðoxa1 Þ ¼ E0 ðox1 Þ\0: ð2:25Þ
In the context of unbiased accounting, we obtain equality.

Indeed, at the time t = 0, the evaluation function becomes:
P0 = fa0 + β1v10 + β2v20, or P0 + d0 = g0 = β1v10 + β2v20; so expectations about
new growth opportunities at time t = 0 depend only on v10 and v20. The relevance
of this reasoning is relevant in emerging sectors, such as the Internet. It is a new
sector in which expectations about high future abnormal returns are sustained in
non-financial variables (vt), whose proxies are associated with web-traffic variables,
such as “number of visitors per website”, “average time spent on website”, “per-
centage of the Internet user population”, etc. (Copeland et al. 2000; Damodaran
2001; Tockic 2004, 2005, among others).
Thus, in a scenario of unbiased accounting and according to LIM:
E0 ðoxa1 Þ ¼ E0 ðox1 Þ ¼ v10 ¼ 0 ðgiven Proposition 9Þ;

E0 ðoa1 Þ ¼ v20 [ 0 ðgiven Proposition 9Þ; ð2:26Þ
E0 ðc1 Þ ¼ v10 v20 ¼ v20 ðsee Eq: 2:20Þ:
The value of investing is v20 (negative values for the variable ct represent
investments in operating assets), which is fully capitalized.
In a scenario of conservatism accounting, and also taking into account
Proposition 9, we obtain:
E0 ðoxa1 Þ ¼ E0 ðox1 Þ ¼ v10 \0;

E0 ðoa1 Þ ¼ v20 [ 0; ð2:27Þ
E0 ðc1 Þ ¼ v10 v20 :
In this context, part of the cash flow is absorbed by losses because the value
invested corresponds to v20, but only the portion (v20 − v10)/v20 < 1 is capitalized,
and the value (v10 < 0) is recognized as (negative) results in the net income
statement (v10 < 0).
Feltham and Ohlson (1995) demonstrated, and considering the dynamic of
information at the time of its initialization, i.e. at t = 0, and assuming that at this
time the goodwill is null, the conservatism accounting effect explains how the first
years of a company’s life can register negative results. However, this situation
begins to reverse itself as the company continues to invest. Based on the principle
of rationality, the company only continues investing if the investments (growth
opportunities) are generating abnormal returns, i.e. Et ðoxat Þ [ 0, because the
objective of managers is to maximize the value of the company or equivalently
maximize the selection of projects with positive NPV.
With reference to business start-ups, particularly companies in technology
industries, the net results are not a good proxy for future results, as they tend to
incorporate high amounts allocated on investments in intangible assets, recorded as

costs, as imposed by GAAP. Core et al. (2003) claim that IPO companies still report
losses, because those companies are holding in their “portfolio” high-growth
opportunities, so investors’ attention focuses on the expectations of future growth.
Myers (1977) has associated the persistence of losses with the existence of a
high-growth opportunities portfolio.
Thus, using the RIV model, it is easy to explain why the multiples of price-to
book value (P/B) and the (PER) tend to be high for those companies due the high
expectations of future growth. Thus, defining the net profit (xt) as xt = ROEt * bvt−1
(return on equity—ROE) and rewriting the RIV model (Eq. 2.6), where r identifies
the cost of capital, we obtain19:
X
1
Et ðROEt þ s bvt þ s1 rbvt þ s1 Þ
MVEt ¼ bvt þ :
s¼1
ð1 þ rÞs
Dividing the above expression by the bvt variable, we obtain the multiple of
book value (P/B):
MVEt X1
Et ½ðROEt þ s rÞbvt þ s1
¼ 1þ : ð2:28Þ
bvt s¼1
ð1 þ rÞs bvt
Equation
2.28 shows
that the P/B ratio is a function of the expected abnormal
Et ðROEt þ s rÞ
results ð1 þ rÞs
and the growth of “stock” needed to generate results in the
future ðbvt þ s1 = bvt Þ. In the absence of abnormal returns the value of this multiple
is the unit.
Calculating now the PER, we add the dividends (dt) to the RIV model, i.e.
P Et ðxatþ s Þ
Pt þ dt ¼ ðbvt þ dt Þ þ 1 s¼1 ð1 þ rÞs . Given the CSR property, we can define bvt +
dt = bvt−1 + xt−1, and replacing this expression in the expression above, we obtain
P Et ðxatþ s Þ
Pt þ dt ¼ ðbvt1 þ xt Þ þ 1 s¼1 ð1 þ rÞs , now calculating the PER we get:
" #
P t þ dt 1þr 1 X 1
Et ðxatþ s Þ xat
¼ þ : ð2:29Þ
xt r xt s¼1 ð1 þ rÞs r
19
Recall that the RIV model is equivalent to the PVED model, which assumes as a theoretical
framework an economy, where the preferences of agents are homogeneous and they are risk
neutral.
2.3 The Effect of Conservatism Accounting 35
FE = CE
FE
P/B ratios higher
PER ratios higher
(FE> CE)
CE
P/B ratios reduced
PER ratios reduced
(FE<CE)
Source: White et al. (1997: 1071) adapted.
Fig. 2.1 The relationship between the multiples P/B and PER and future abnormal earnings (FE)
and current earnings (CE)
The PER summarizes the determinant of the magnitude of this

ratio, i.e. the
P1 Et ðxatþ s Þ
difference between future abnormal earnings s¼1 ð1 þ rÞs FE and current
a
xt
abnormal earnings r CE . Graphically this is expressed in Fig. 2.1.
Along the 45° line, the expected abnormal earnings (FE) equals current earnings
(CE). In this context, the current results are a good indicator of future performance.
Above this line, growth affects PER ratios and future abnormal earnings
(FE) exceed the current earnings (CE). Note that, even when the net results are
negative, the PER ratios can take high values. In this context, the negative results
are seen as transitory, e.g. the result of large investments in intangibles, recorded as
costs, particularly in technology-based companies in the start-up phase (accounting
conservatism).
Below the 45° line, because the future earnings (FE) are lower than the current
earnings (CE), low PERs appear because the high values of CE are due to the
presence of transitory items, which are not expected to persist in the future.
In summary, goodwill, measured by the difference between the market value
(MVE) and the book value of equity (BVE) of the company, is the result of a dual
effect: (i) the undervaluation of assets (conservatism accounting); and (ii) overes-
timation of the expected abnormal results. These effects are more pronounced in
technology-based companies, especially in the start-up phase. Therefore, in
accordance with LIM, the FOM at the time of the establishment of the company and
taking into account Proposition 9 as the same model, show that only part of the
investment is capitalized, with the remainder recognized as cost. However,
assuming the principle of rationality, it is expected that managers invest in new
growth opportunities if they are associated to abnormal returns. Hence, the net firms
and companies with contemporaneous initial purchase offer (IPO) dates (including
non-net firms and very young companies during the start-up phase), often propose
a new untested business idea), “moving” to the market with high losses valued
positively by the market and sustain the phenomenon of negative pricing of losses,
i.e. the positive valuation of losses. The evaluation of this type of company is
particularly difficult, given the difficulty of estimating future cash flows. The
volatility of their prices reflects this phenomenon.
Appendix 2.1: Deduction of the OM Model

(Eqs. 2.8 and 2.9)
1. Deduction of the Ohlson (OM) (1995) model—Eq. 2.8:

Ohlson (1995) defines the valuation function as:
Pt ¼ bvt þ a1 xat þ a2 vt :
with the parameters assuming the values:

w
a1 ¼ 0 and
Rf w
Rf
a2 ¼ 0:
ðRf wÞðRf cÞ
To obtain this function we first assume:

w 1
(i) The matrix P ¼ R1 ;
f 0 c
(ii) The
a dynamic of information is defined as:
xt þ 1 ; vt þ 1 ¼ Rf P½xa;
t ; vt þ e1;t þ 1 ; e2;t þ 1 and,
(iii) Assuming that the supranormal results are defined as:
a
Rs s a
f Et xt þ s ¼ ½1; 0P xt ; vt :
Based on the Residual Income Valuation Model (RIV), the expression assumes:
X
1 a a
Pt bvt ¼ Rs
f Et xt þ s ¼ ½1; 0 P þ P þ . . .
2
xt ; vt :
s¼1
Appendix 2.1: Deduction of the OM Model … 37
A series of matrices P + P2 + … are convergent, given the square root of the

characteristic of the matrix is less than unit. Thus, we could conclude that:
½1; 0P½I P1 ¼ ½a1 ; a2 ; i:e::
1
1 w 1 1 0 1 w 1
½1; 0 Rf Rf ¼ ½ a1 ; a2
0 c 0 1 0 c
" #1
1 1
1 R1 f w R1
Rf w Rf f
¼ ½ a1 ; a2
0 1 R1 f c
2 R1
3
1
1
f
Rf w R1 4 f1R 1 w
ð f Þð f Þ 5 ¼ ½ a ; a
1R1 w 1R 1 c
f 1 2
1
0 1R1 c
f

R1
f w R1
f wRf
1
R1
þ f
¼ a a
1R1f
w ð f Þð f Þ
1R 1 w 1R 1 c 1R 1 c
f
1; 2
8 1 1
> R w
< a1 ¼ 1Rf 1 w ¼ R1 ðR
R f w
¼ Rfww
f f f wÞ
>
: a2 ¼ R1
f wRf
1
R1 R1 wR1 R1
1 c þ ¼ R1 ðRf wÞR1 ðR cÞ þ
f f f f
ð1R1 w Þ ð 1R Þ 1R1
f
c f f f R1
f
ðRf cÞ
(
f f
a1 ¼ Rf w
w
Rf w :
a2 ¼ ðRf wÞðR
w
f cÞ
þ ðRf wÞðRf cÞ ¼ ðRf wÞðR
Rf
f cÞ
2. Deduction of the Eq. 2.9 of the OM Model

Assuming the Eq. 2.8:
Pt ¼ bvt þ a1 xat þ a2 vt ;
and substituting xat , we obtain,

Pt ¼ yt þ a1 ½xt ðRf 1Þyt1 þ a2 vt :
Due the CSR principle:
Pt ¼ bvt þ a1 xt a1 ðRf 1Þðbvt xt þ dt Þ þ a2 vt ,

Pt ¼ bvt þ a1 xt a1 ðRf 1Þbvt þ a1 ðRf 1Þxt a1 ðRf 1Þdt þ a2 vt ,;
Pt ¼ bvt ½1 a1 ðRf 1Þ þ a2 vt þ a1 xt þ a1 ðRf 1Þxt a1 ðRf 1Þdt and j ¼

a1 ðRf 1Þ thus,
Pt ¼ bvt ð1 kÞ þ a2 vt þ a1 ½xt þ ðRf 1Þxt ðRf 1Þdt ,
Pt ¼ ð1 kÞbvt þ a2 vt þ a1 Rf xt a1 ðRf 1Þdt , and u ¼ RfR1f
, we obtain:
Pt ¼ jðuxt dt Þ þ ð1 jÞbvt þ a2 vt :
Appendix 2.2: Deduction of the Equivalence

of the Preposition Number 1 of Feltham and Ohlson
(FOM) (1995) Model and Eqs. 2.15a, 2.15b and 2.15c
P
(i) Equation 2.15a: Pt ¼ fat þ 1 s
s¼1 Rf Et ðct þ s Þ
Assuming the FAR [fat = fat−1 + it − (dt − ct)] and the definition of it
[it = (Rf − 1)fat−1], the dividends could be expressed as:
dt ¼ fat1 þ it þ ct fat ¼ fat1 þ ðRf 1Þfat1 þ ct fat

¼ Rf fat1 þ ct fat:
Hence for any sequence of the variables cte fat ({ct+τ, fat+τ}τ≥1), valuation
function is:
X
1 X
1
Pt ¼ Rs
f Et ðdt þ s Þ ¼ Rs
f E t ½Rf fat þ 1s þ ct þ s fat þ s
s¼1 s¼1
X
1
¼ fat þ Rs
f Et ðct þ s Þ;
s¼1
because Rsf Et ðfat þ s Þ ! 0 com s ! 1.

P a
(ii) Equation 2.15b: Pt ¼ bvt þ 1 s
s¼1 Rf Et xt þ s
This equations corresponds to the Residual Income valuation Model defines
in Eq. 2.6. a
P
(iii) Equation 2.15c: Pt ¼ bvt þ 1 s
s¼1 Rf Et oxt þ s
Assuming operational and non-operational activities, the following expres-
sion corresponds to the operational abnormal results: oxat ¼ oxt ðRf 1Þ
oat1 , applying the OAR we obtain:
oat ¼ oat1 þ oxt ct ; substituting the variable oxt

ct ¼ oat1 þ oxat þ ðRf 1Þoat1 oat ,
ct ¼ oxat þ Rf oat1 oat :

For any sequences of the variables oxatþ s ; oat þ s1 gs 1 , we obtain:
X
1 X
1 a X
1 a
Rs
f Et ðct þ s Þ ¼ Rs
f Et oxt þ s þ Rf oat1 þ s oat þ s ¼ oat þ Rs
f Et oxt þ s ;
s¼1 s¼1 s¼1
thus, Rs
f Et ðoat þ s Þ ! 0 com s ! 1:
Appendix 2.2: Deduction of the Equivalence of the Preposition … 39
If we add fat to the above expression, and due the Eq. 2.15a, the valuation
function is:
X
1 X
1 X
1
Pt ¼ Rs
f E t ðdt þ s Þ ¼ ðoat þ fat Þ þ oxatþ s ¼ bvt þ oxatþ s :
s¼1 s¼1 s¼1
Appendix 2.3: Deduction of the Feltham and Ohlson (1995)

—Eq. 2.22
Assuming the definition of goodwill:
gt ¼ Pt bvt
Multiplying the above expression by Rf and according Ohlson (2000):

Rf gt ¼ Et gt þ 1 þ oxatþ 1 :
gt could be defined as a linear function, such as:
Pt bvt ¼ gt ¼ a1 oxat þ a2 oat þ b vt
with

Rf gt ¼ Rf a1 oxat þ a2 oat þ b vt ¼ Et a1 oxatþ 1 þ a2 oat þ b vt þ oxatþ 1 ,

Rf a1 oxat þ Rf a2 oat þ Rf b vt ¼ Et ða1 þ 1Þoxatþ 1 þ a2 oat þ 1 þ bvt þ 1 :
Due the structure of Linear Information Model (LIM):
Rf a1 oxat þ Rf a2 oat þ Rf b1 v1t þ Rf b2 v2t ¼ ða1 þ 1ÞEt ðoxatþ 1 Þ þ a2 Et ðoat þ 1 Þ þ bEt ðvt þ 1 Þ
Substituting the expected values according the dynamic information:
Rf a1 oxat þ Rf a2 oat þ Rf b1 v1t þ Rf b2 v2t ¼ ða1 þ 1Þðw11 oxat þ w12 oat þ v1t Þ
þ a2 ðw22 oat þ v2t Þ þ b1 c1 v1t þ b2 c2 v2t
Solving the equation based on that the probability should be one, thus:
8 8 8
> Rf a1 ¼ ða1 þ 1Þw11 > Rf a1 ¼ a1 w11 þ w11 > a1 ðRf w11 Þ ¼ w11
>
> >
> >
>
< R a ¼ ða þ 1Þw þ a w < R a a w ¼ ða þ 1Þw < a ðR w Þ ¼ ða þ 1Þw
f 2 1 12 2 22 f 2 2 22 1 12 2 f 22 1 12
, ,
>
> Rf b1 ¼ ða1 þ 1Þ þ b1 c1 >
> Rf b1 b1 c1 ¼ ða1 þ 1Þ >
> b1 ðRf c1 Þ ¼ ða1 þ 1Þ
>
: >
: >
:
Rf b2 ¼ a2 þ b2 c2 Rf b2 b2 c2 ¼ a2 b2 ðRf c2 Þ ¼ a2
8 8
a1 ¼ Rf w11
w11 > a ¼ w 11
>
> >
>
1 Rf w11
>
< a ðR w Þ ¼ ð w11 þ 1Þw >
< a2 ðRf w22 Þ ¼ ðw11 þ Rf w11 Þw12
2 f 22 Rf w11 12 Rf w11
, ,
>
> b ðR c Þ ¼ w11
þ 1 >
> b ðR c Þ ¼ w11 þ Rf w11
>
:
1 f 1 Rf w 11 >
> 1 f 1 Rf w11
:
b2 ðRf c2 Þ ¼ a2 b2 ¼ Rfac 2
8 2
>
> a1 ¼ Rfww 11
>
>
11
>
< a2 ¼ ðR wRfÞwðR12 w Þ
, :
f 11 f 22
>
>
R
b1 ¼ ðRf w11 ÞfðRf c Þ
>
>
>
: b ¼ a2
1
2 Rf c
2
References
Barth M, Cram D, Nelson K (2001) Accruals and the prediction of future cash-flows. Acc Rev 7
(1):27–58
Beaver WH (2002) Perspectives on recent capital research. Acc Rev 77(2):453–474
Copeland T, Koller T, Murrin J (2000) Valuation: measuring and managing the value of
companies, 3rd edn. Wiley, New York
Core JE, GuayWR, Buskirk AV (2003) Market valuations in the new economy: an investigation of
what has changed? J Acc Econ 34(1, 3):43–67
Damodaran A (2001) The dark side of valuation. Prentice Hall, US
Edwards EO, Bell PW (1961) The theory of and measurement of business income. University of
California Press, Berkeley, Los Angeles
Fama E (1976) Foundations of finance. Basic Books, New York
Feltham G, Ohlson J (1995) Valuation and clean surplus accounting for operating and financial
Feltham G, Ohlson J (1999) Residual earnings valuation with risk and stochastic interest rates. Acc
Rev 74:165–183
Ferreira L, Sarmento M (2004) Que modelos adoptam os Revisores Oficiais de Contas na
Avaliação de Empresas e de Participações Sociais? Revisores e Empresas 6(25):19–31
Gordon M, Shapiro E (1956) Capital equipment analysis: the required rate of profit. Manage Sci
3:102–110
Holthausen RW, Watts RL (2001) The relevance of the value-relevance literature for financial
accounting standard setting. J Acc Econ 31:3–75
Jensen M (1986) Agency costs of free cash flow, corporate finance and takeovers. Am Econ Rev
76(2):383–393
Kothari SP (2001) Capital markets research in accounting. J Acct Econ 31(1,3):105–231
Lo K, Lys T (2001) The Ohlson model: contribution to valuation theory, limitations, and empirical
applications. J Acc Audit Financ 15(3):337–367
McCrae M, Nilsson H (2001) The explanatory and predictive power of different specifications of
the Ohlson, J. (1995) valuation models. Eur Acc Rev 10(2):315–341
References 41
(4):411–433
Modigliani F, Miller M (1958) The cost of capital, corporate finance and the theory of investment.
Am Econ Rev 48(3):261–297
Morgenstern O (1963) On the accuracy of economic observations, 2nd edn. Princeton University
Press, Princeton
Myers SC (1977) Determinants of corporate borrowing. J Financ Econ 5:147–175
(2):661–687
Ohlson J (2000) Earnings, book values, and dividends in equity valuation: an empirical
perspective. Contemp Acc Rev 18(1):107–120
Rayn S (1995) A model of accrual measurement with implications for future profitability, risk and
market value. Rev Acc Stud 33:95–112
Richardson G, Tinaikar S (2004) Accounting based valuation models: what have we learned? Acc
Financ 44(2):223–333
Stober TL (1999) Empirical applications of the Ohlson [1995] and Feltham and Ohlson [1995,
1996] valuation models. Manag Financ 25(12):3–16
Tokic D (2004) R&D, advertising and the market value of internet firms: part I. J Internet Commer
3(2):79–99. Doi:10.1300/J179v03n02_05
Commer 4(4):23–40. Doi:10.1300/J179v04n04_02
White GI, Sondhi AC, Fried D (1997) The analysis and use of financial statements, 2nd edn.
Wiley, US
Williams JB (1938) The theory of investment value. Harvard University Press, Cambridge
Zhang X (2000) Conservative accounting and equity valuation. J Acc Econ 29(1):125–149
Chapter 3
The Value Relevance of the Variables
Earnings and Book Value of Equity
for Valuation Purposes
Abstract Since the seminal work of Ball and Brown (J Acc Res 6(2):159–1478,
1968) and Beaver (J Acc Res 6:179–192, 1968), extensive literature has examined
the relationship between the prices of securities and (positive) results reported by
companies. However, analysis of price/losses is scarce, with contradictory results
obtained so far. Ratio (price versus losses) gained relevance in 1990s, with
well-documented results in new economy companies (Shiller in Irrational exuber-
ance. Princeton University Press, Princeton, 2000). It is for this group of companies
that the phenomenon “positive valuation losses” assumes its great of significance,
although it was not an entirely new phenomenon. For example, Amir and Lev (J
Acc Econ 22(1–3):3–30, 1996) documented this relationship in the mobile phone
sector, also with reference to the US market, in the 1980s.

Keywords New economy companies Positive valuation of losses Value rele-
vance of earnings and book value of equity
It is important to point out that this effect of positive valuation of losses occurred in
a decade when there was an increasing number of small and young businesses,
particularly technology-based firms (Hayn 1995; Basu 1997; Burgstahler and
Dichev 1997; Collins et al. 1997, 1999; Chan et al. 2001; McCallig 2004; Joos and
Plesko 2005) which report after the 90s, a greater magnitude of losses and a
tendency to persist for longer periods of time (McCallig 2004; Joos and Plesko
2005). These effects, i.e. the report of losses and its persistence over time are
particularly dominant in the samples analyzed (see Tables 5.3 and 5.4).
In this chapter, we will begin by analyzing the information content of the
variable results for evaluation purposes. Briefly, we begin by analyzing profitable
firms, and then focus our analysis on companies with losses. After that, we will
introduce into the analysis the variable BVE which, along with the variable results,
net income, is the main determinant of the value according to the OM and FOM, the
theoretical framework in this study, and based on which we intend to empirically

DOI 10.1007/978-981-10-3009-3_3
44 3 The Value Relevance of the Variables Earnings …
test the phenomenon of positive valuation of losses. As stated by Ohlson and

Penman (1992: 553):
These two summary measures (“book value of equity” and “earnings”) achieve pre-eminent
status by serving as primary indictors of a firm’s value.
3.1 The Informational Content of the Variable Results
The variable results is relevant for evaluation purposes, or in other words, investors
see the variable results as a valuable source of information linked to the company’s
value. According to the model of discounted cash flows, the value of a share results
from the discounted future cash flows (CF), and the discount rate reflects the
opportunity cost of capital and the risk inherent to the business (r). Assuming a flow
of constant returns with an infinite duration (perpetuity), and based on the principle
of continuity (going concern), the price [numerical expression of the value (Neves
2002: 4)] corresponds to:
E0 ðCFÞ E1 ðCFÞ X1
E0 ðCFt Þ CF
P0 ¼ þ ¼ ¼ t ¼ ð3:1Þ
ð1 þ rÞ ð1 þ rÞ 2
t¼1 ð1 þ rÞ
r
Considering the variable results (X) as a proxy for future cash flows (unob-
servable) (Modigliani and Miller 1966), the price is as follows1:
X
P0 ¼ ð3:2Þ
r
In this context, there are multiple studies that aim to examine the relationship
between the price movements and the results reported by companies, i.e.:
Pi ¼ a þ bXi þ ei ð3:3Þ
If the variable results are relevant for evaluation purposes, it is expected that the
parameter β should be positive and statistically significant because it measures the
covariance between the variable price and results corresponding to the multiple of
the results (PER = P0/X = 1/r). However, empirical studies have documented for
the parameter β low values ranging from one to three (Kormendi and Lipe 1987)
1
White et al. (1997: 1043) and Ferreira and Sarmento (2004: 31) show, in theory, the equivalence
between the models based on the discount of cash flows and models based on stock of assets held
by the company, from which is generated the expected revenue. If we assume that CF = ROE *
(BVE) and in the medium–long term the ROE = r, being r the rate of the cost of capital and ROE
the rate of return on equity, then: MVE = CF/r = (ROE * BVE)/r = BVE. Thus, as shown by
Ohlson (1995), over the medium–long term the variable BVE is an unbiased estimator of MVE.
3.1 The Informational Content of the Variable Results 45
when theoretically the values of β, assuming a reasonable estimate for the PER
ratio, should be between 8 and 20 (Kothari 2001: 129). Thus, even assuming that all
companies record profits, the explanatory power of this model already shows very
low values.2
There are several explanations to support the weak explanatory power of the
variable results. Beaver et al. (1980) justify the lack of synchronization between the
movements of prices and results based on the gap between the arrival of the new
information on the market, since they are incorporated into the prices immediately
but only later reported in financial statements [an effect captured by Ohlson (1995)
and Feltham and Ohlson (1995) by the variable vt]. Thus, there is increasing
pressure for solid reporting of information from the company, i.e. for the company
to report more and more information, also extra financial information, in particular
information relating to business risk, and in a timelier manner. Guimarães (2004)
suggests that the Internet is a very valuable vehicle for online dissemination of
information.3
The “contamination” of the variable results with transitory items, which intro-
duces noise into the variable results, is another explanation provided by Collins
et al. (1994) and Ramakrishnan and Thomas (1993).4 Ohlson and Shroff (1992) and
Kothari and Zimmerman (1995) postulate that econometric factors justify the weak
explanatory power between prices and results. They demonstrate that the choice of
the methodology price level (Eq. 3.3), which assumes that results follow a random
walk5 enables prices to reflect a broader and timely set of information, provides
2
If the variable results constitute an adequate proxy for future cash flows, the cost of capital is
identical for all companies, and assuming no growth, the variable “results” should explain 100% of
the cross-section changes in the variable “prices”. For a detailed analysis see Kothari (2001).
3
The importance of standardization of financial reporting is such that the new XBRL Extensible
Business Reporting Language (XBRL) is being used for the electronic communication of business
and financial data, cause of its versatility to be used throughout the world, regardless of language,
type of business or adopted accounting system. For example, the Security Exchange Commission
(SEC) announced that it will engage directly in the development of financial reporting globally,
like the International Accounting Standard Board (IASB).
4
It is assumed that the variable “results” (X) has three components: (i) permanent results (XP), i.e. a
flow of results that that the company is expected to sustain in the future, whose “persistence”
depends among other factors of the customer portfolio held by the company, such as customer
loyalty to the brand and product quality; (ii) temporary items (XT), potentially resulting from
discontinued operations and extraordinary items, and (iii) irrelevant items for valuation purposes
(XO) resulting for example from changes in accounting policies. Analytically, the price (Pi) is a
function of Pi = α + βPXP + βT XT + βo XO. Theoretically, the parameter βT must take the value one
(βT = 1), because it is expected that the impact of a monetary unit of a transitory item in results has
a similar impact on prices, restricting their effect to the year in which it is reported. ΒP > βT,
because the market assigns a higher multiple (PER) to the permanent results given its persistence,
which is expected to continue in the future. The parameter β0 assumes a null value, given the
irrelevance of this variable for evaluation purposes.
5
The assumption that the results follow a random walk is sustained in the research of Ball and
Watts (1972). But while the assumption that the prices follows a random walk is based on a solid
body of theory, the theory of efficient markets, the application of this assumption to the variable
‘results’ is merely indicative.
non-biassed coefficients compared to the return models and differentiated-price

models, which tend to be more used because they circumvent econometric problems
such as heteroscedasticity.6 However, these authors found that different method-
ologies continue to predict the constant statistically significant values different from
zero, which is inconsistent with the theory. Thus they recommend the use of
different methodologies, suggesting that the potential existence of nonlinear rela-
tionships between prices and the variable results should be explored (either
empirically or theoretically).
Easton et al. (1992) in turn, invoke the clean surplus principle, i.e. that all
available information sooner or later is incorporated into the variable results, thus
they suggest that in returns models the investigations should cover a more broader
horizons for measuring the variable results. The aim is to reduce potential errors of
measurement of this variable. The empirical results obtained by these authors
confirm the existence of a greater correlation between the variables returns and
results when expanding the time horizon to measure the variable results.
The existence of a negative correlation between the free-risk interest rate and the
parameter β (PER) is suggested by Collins and Kothari (1989). As the discount rate
is calculated by adding a risk premium to the risk-free rate, the increase in the
interest rate without risk (ceteris paribus) leads to the parameter β facing a decrease
in value. This negative relationship stems from the (unexpected) impact of inflation
on economic activity: the variation in the free-risk interest rates results from vari-
ations in the inflation rate, and, if the company reflects this effect in an increase in
prices of goods/services sold, it is not expected that there is link between the
risk-free rate and the parameter β (PER). Table 3.1 summarizes the main studies.
More recently, and due the greater number of companies registering losses, a line
of research argues that the information content of the variable results for evaluation
purposes is asymmetric, i.e. depending on the company reporting profits or losses.
The asymmetry derives from the liquidation option that shareholders hold from the
company. Traditionally empirical studies removed from the studies companies with
losses (e.g. Beaver et al. 1980; Collins and Kothari 1989).
One of the pioneering studies in this area was the study of Hayn (1995). Hayn
(1995) uses the returns model, where the annual returns are explained in terms of
the annual results by shares to test the hypothesis
bðand R2 )companies with losses \bðand R2 )global sample\bðand R2 )profitable companies:
6
In the price level models, the variable prices are explained in terms of earnings per share. In return
models, performance is explained by the earnings per share, deflated by the price at the beginning
of the period. In turn, the differentiated-price models consider variations in either returns or results
(Donnelly 2002). Donnelly (2002) shows similar results for the three methodologies if the variable
results are not “contaminated” by transitory items. In the presence of these items, the results of
different models tend to differ dramatically depending on the methodology used.
Table 3.1 Summary of the factors supporting the weak correlation between the variable results
(positive) and the movement of prices
Authors Explanatory facts
• Beaver et al. (1980) – Lack of timeliness effect associated with financial information
• Ramakrishnan and – The noise introduced by transitory items (item 17 of the annual
Thomas (1993) Compustat database) in the variable results
• Collins et al. (1994)
• Ohlson and Shroff – Econometric factors. The different methodologies (price level
(1992) regressions, returns model and different-priced models) tend to
• Kothari and produce different results if the variable results are contaminated
Zimmerman (1995) by transitory items
• Donnelly (2002)
• Easton et al. (1992) – Need to consider higher horizons to measure the variable results
under the assumption of CSR (clean surplus relation)
• Collins and Kothari – The impact of the free-risk interest rate
(1989)
The results for the period from 1963 to 1990 confirm that the β parameter (and
R2) assumes values near zero in the sample of companies with losses; in samples
with profits the parameter β takes values close to three and the R2 stands at 17%. To
ensure greater robustness of the results, Hayn (1995) creates portfolios based on the
number of consecutive years in which companies report losses; she finds that the
probability of a company reporting losses in a year is 14%, decreasing to eight per
cent in companies that have registered six (or more) years of consecutive losses. As
we increase the number of years in which companies report losses, the parameter β
is negative (positive phenomenon of valuation losses) and R2 even records a slight
increase, which Hayn (1995) justifies on the basis that it is not sustainable that the
company will stay in the market, registering consecutive losses, because in this
context shareholders will liquidate the company.
When a company reports losses in a year, the price does not decrease necessarily
to zero, but in proportion to the variation in the variable results. The justification for
such behaviour, and according to Hayn (1995), results from the fact that the market
interprets the losses as temporary, given the liquidation option held by shareholders.
As shareholders hold a put option on future cash flows of the company, if the
company faces the losses or unsatisfactory levels of results, shareholders exercise
the liquidation option and sell their shares at a price proportionate to the liquidation
value of the assets. Thus Hayn (1995) demonstrates positive valuation of losses (see
Fig. 3.1).
With high levels of results (X), the value of the company results from capital-
ization of the results (k = PER). Reporting losses or unsatisfactory levels of results,
the company value is V = L, where L is the liquidation value of the company.
X* = L/k is the level of results below which it is preferable to shareholders exercise
V=kX
V – company value
V=L
(0.0) X*=L/k
X - results
Source: Hayn (1995: 133)
Fig. 3.1 The relationship between the value of the company’s equity and results due to the
liquidation option held by shareholders
the liquidation option.7 Thus, Hayn (1995) shows that, for profitable companies
(X > X*), the price is strongly correlated with the results, tripling the multiple
assigned by the market to such companies as seen, this variable acts as an appro-
priate proxy for expected future income streams as postulated by Modigliani and
Miller (1966). For companies reporting losses, or with unsatisfactory levels of
results (X < X*—temporarily depressed firms), there seems to be no correlation
between the price and the results (losses) because, in the case of persistent losses,
shareholders opt to liquidate the company. These results are consistent with the
existence of a nonlinear relationship between the variables “price” and “results”,
when the latter reaches extreme values.
Similarly, where the losses registered by a company tend to be transitory and so
are irrelevant for assessment purposes, Subramanyam and Wild (1996) document
an inverse relationship between the probability of default and persistence of the
variable results, i.e. losses. The authors begin by estimating the returns obtained by
investors on the basis of surprises registered at the level of results (unexpected
earnings). Later, the probability of insolvency is introduced into the analysis, based
on the methodology proposed by Altman (1968), as an interactive variable (i.e.
calculating the product of the variable “surprises” in terms of results—unexpected
7
Empirically Hayn (1995) defines two proxies to the liquidation value. The first proxy is identified
with the credit risk rating assigned to the company, according to the methodology of Standard &
Poor’s. The second proxy is defined as follows [(P-L)/PEIND]/σPER, where P is the market value of
the equity of the company, L is the liquidation value of the company, PEIND is the average PER
ratio for the industry, and σPER corresponds to the standard deviation of the PER ratio of the
company. The numerator expresses the excess market value of the company in results-based terms.
The denominator measures the magnitude of the need to decrease the value of equity to be equal to
its liquidation value.
results).8 As expected, the informational content of the variable results decreases as

the losses become persistent because of the greater probability of insolvency of
these companies. The results for the period 1981–1990 seem robust to the intro-
duction of control variables that aim to capture the effect of: (i) failure of the debt
service; (ii) the variability of the variable results (risk); and (iii) the correlation
between the increase in the number of losses and the probability of insolvency.
Chambers (1997) shows similar results to Subramanyam and Wild (1996), i.e.
the variable “results”, when negative (losses), are irrelevant for valuation purposes.
Chambers (1997) argues that the weak explanatory power of the variable “losses”
depends on their degree of persistence (measured by the ratio between the present
value of accumulated losses and the first loss reported). Subdividing the sample into
several groups depending on the level of persistent losses, Chambers (1997) finds
that, for the groups in which losses tend to be more persistent, the information
content of the variable initial loss is more significant. These results show that faced
by the persistence of losses, investors immediately incorporate the greatest degree
of bankruptcy which these companies incur, a fact that, according to Chambers
(1997) justifies the weak explanatory power associated with variable results when
successive losses are reported because in this context the exercise of the liquidation
option is imminent.9
Basu (1997) follows Hayn (1995), Subramanyam and Wild (1996) and
Chambers (1997) line of research, to explore the existence of an asymmetric
relation between the movements of the variable “prices” and the changes in the
variable results. However, the theoretical basis is different. For Basu (1997),
asymmetry in the information content of the variable results is due to the effect of
the principle of prudence (one of the four types of accounting conservatism iden-
tified by Richardson and Tinaikar (2004)—see Sect. 2.3). According the principle
of prudence in preparing the financial statements, actual and potential losses should
be recognized immediately, while the potential gains should be deferred, i.e. they
are measured only when realized: as a result, the variable results record losses more
timely when compared to gains. To empirically measure this effect, he adopts a
different method compared to Hayn (1995). Thus, in line with Beaver et al. (1980),
and given that prices reflect a wider range of information than the financial infor-
mation, Basu (1997) adopts the reverse regression methodology, where the annual
results are now explained on the basis of annual returns. Assuming that the capital
8
The bankruptcy probability (Zit) is calculated based on the methodology of Altman (1968) and is
defined as Zit ¼ 1:2X1;it þ 1:4X2;it þ 3:3X3;it þ 0:6X4;it þ 1:0X5;it , where (X1) measures the ratio
of the difference between the value of current assets and current liabilities to total assets; (X2)
expresses the proportion of retained earnings over total assets; (X3) measures the results before
interest and taxes (EBIT) over total assets; (X4) is the ratio of the market value of equity and
preferred shares to the book value of total debt; and (X5) is the quotient between total sales over
the total assets. As Zit is a continuous variable, the shorter the value of Zit, the higher the
probability of the company becoming insolvent.
9
Note that Chambers (1997) excludes the sample companies in the start-up phase for the period
1976–1992 to avoid potential bias in the results resulting from the growth effect.
(a) (b) Rit - Returns
I
II I
II
Xit - Profits
Rit - Returns
III
IV IV
III
Fonte: Basu (1997: 12).
Fig. 3.2 a Relationship between the variables profits and returns according the principle of
prudence. b Relationship between the variables profits and returns given the liquidation option by
shareholders
market is perfect, the variable “returns” acts as a proxy for good and bad news.10 As
expected, the parameters associated with the recorded negative returns have higher
statistically significant values compared to the positive returns. Basu (1997) con-
cludes that, in being recognized immediately, the effect of the losses is limited to
the period in which they occur, whereas future results are protected from bad
current news. In contrast, profits tend to be more persistent and last over time, as
they are recognized only when effective.
Establishing a comparative analysis with the results of Hayn (1995), Basu (1997)
concludes that Fig. 3.2a (his own) is identified with Fig. 3.2b Hayn (1995) along a
45° reversed line. In Fig. 3.2a, with the use of reverse regression methodology, the
parameter β is higher for companies that reported losses because, in accordance
with the principle of prudence, the actual and/or potential losses are recognized
immediately, while profits (potential) are deferred. Using the traditional method, in
which the results (unexpected earnings) are the explanatory variable, they have a
higher magnitude for the parameter β (Fig. 3.2b), confirming that for evaluation
purposes, the market assigns high value to multiples because the variable results are
seen as a consistent proxy for future results. As for the constant, Basu (1997) justify
10
The empirical model estimated by Basu (1997) corresponds to Xit =Pi;t1 ¼ a0 þ a1 DRit þ
a2 Rit þ a3 Rit DRit , where Xit corresponds to the results of the company i for the period t, Pi, t−1
is the price of the shares i at beginning of period t (end of t − 1), DR is a dummy variable that
takes the value 1 when the company reports negative results (bad news), zero otherwise.
According to Basu (1997), considering the variable “price”, based on which are calculated
returns (R) as an independent variable, and given that this variable reflects all available infor-
mation in a more timely manner, the estimates provided by the method OLS (Ordinary Least
Squares) are more accurate.
contrary to Hayn (1995) that positive values are a reflection of the good news of
previous periods whose gains were deferred and only now (when effective) are
recognized.
In this context, Basu (1997), along with Hayn (1995), Collins et al. (1997, 1999),
McCallig (2004) and Joos and Plesko (2005), recognizes that in the period 1963–
1990, there were an increasing number of companies with losses. The increase in
losses derives from a double effect: (i) a greater degree of accounting conservatism in
the reporting of financial information and, (ii) an increase in the liability of auditors.
In compliance with the principle of prudence, auditing the accounts constitutes a
form of control and monitoring by shareholders of the behaviour of managers. In a
context of uncertainty, managers have inside information about the company’s
future prospects. If the managers’ incentive system is indexed to the results, they are
encouraged to adopt accounting policies (creative accounting) that encourage in the
increase of short-term net income. Lenders also claim a more timely reporting of
losses (actual and potential), as the value of the option they hold about the company
(put option) is more sensitive to decrease, rather than to increase, the value of the
company. These requirements are particularly relevant in technology-based com-
panies, whose value is mainly concentrated in intangible assets. In the event of
business failure, such assets come to lose all of their value. Lenders have adopted
the principle of prudence, as a strategy to reduce contract costs (monitoring costs)
(Jensen and Meckling 1976).
The increase in the responsibility of auditors, particularly in view of recent
accounting scandals, is another factor that has boosted a greater degree of
accounting conservatism in the reporting of financial information. Basu (1997) also
notes that taxes and the adoption of regulatory policies as factors intended to
standardize accounting practices and procedures in order to achieve comparability
between financial statements.11
In summary, and based on the above studies, we conclude that losses are
irrelevant when estimating future income streams of a company or the market value
of equity (MVE) because, if they exist, they tend to be transitory. If they do persist,
then shareholders exercise the liquidation option they hold on the company’s assets.
These results raise the question about the homogeneous relationship that was
assumed to exist between the movement of the variables “prices” and “results”.12 In
this context, the variable (BVE) is a relevant determinant of the value of a company
according OM and FOM.
11
The standardization of accounting practices and policies is a global trend. As an example, we
refer to the Regulation of the European Community (EC) no. 1606/2002 of the Parliament and of
the Council of 19 July 2002 on the application of International Accounting Standards (IAS), in
order to ensure a greater degree of transparency and comparability of financial statements with a
view to a more efficient functioning of increasingly global capital markets.
12
Hayn (1995) also notes that multiple results (PER), even for the group of profitable firms, has
been decreasing over time, but she does not suggest any explanation for this result.
3.2 The Informational Content of the Variable Book

Value of Equity
For Collins et al. (1999) the relevance of the variable BVE for evaluation is sup-
ported by three factors: (i) econometrics factors, because the omission of this
variable induces in an estimated bias coefficient for the variable “results”; moreover
the inclusion of variable “BVE” in the valuation model enables control of the effect
of scale/size in “price levels models (Eq. 3.3), as shown by Barth and Kallapur
(1996); (ii) with reference to the OM, the variable BVE is a proxy for future
(normal) expected results; and (iii) the variable BVE is also a relevant proxy for the
liquidation value of the company, when the company is in a situation of financial
stress—for example, reporting continuous losses, as sustained by Hayn (1995),
Berger et al. (1996), Burgstahler and Dichev (1997), Schnusenberg and Skantz
(1998), and Barth et al. (1998).
The variable BVE is particularly relevant for assessment in a context where the
company report losses. In this case, the variable loss is not relevant in predicting the
future profitability of the company, as the losses are not sustainable indefinitely;
then according to the principle of continuity, the BVE is assumed as a proxy for the
present value of future normal results. In the case of persistent losses, the variable
BVE assumes relevance as a proxy for the value of liquidation/abandonment of the
company (e.g. Hayn 1995; Chambers 1997; Subramanyam and Wild 1996). Hence,
the theory shows that the relevance of the BVE variable for evaluation purposes is
inversely related to the financial health of the company (Beaver 2002).
Burgstahler and Dichev (1997) find that both the results and the BVE variables
are relevant for evaluation. However, contrary to the OM, a convex relationship is
sustained between the variables BVE and results—losses. Measuring the BVE, the
level of resources held by the company, it acts as a proxy for the adaptation value of
the company.13 The results in turn act as a proxy for recursion value, i.e. for the
present value of expected cash flows. In this context, the value of the option of
adjustment/liquidation depends on the recursion value, and vice versa. Therefore,
whenever the results reach satisfactory levels (higher return compared to the cost of
capital), the results are the determinant variable for evaluation purposes. When
profitability degraded, the ROE rate is at unsatisfactory levels [e.g. the temporarily
13
Burgstahler and Dichev (1997) choose the name “adaptation option” because only 1% of firms
on the Compustat database are settled. The authors argue that in situations of poor performance,
companies choose to convert their resources, their line (s) of business, to more profitable activities,
carrying out: restructuring, divestitures or investments in new lines of business, mergers and
acquisitions (M&As), takeovers (acquisition of one company by another), spin-offs (partial seg-
regation of an activity), management buyouts (MBO) (purchase of shares by the directors),
sell-offs (expression assigned to a market or segment whose market prices at any given time, are on
a downward trend, usually as a result of a shortage of purchase orders). Those options as suggested
by Wysocki (1998) include the option of exploring new growth opportunities.
3.2 The Informational Content of the Variable Book Value of Equity 53
depressed firms identified by Hayn (1995: 145)]; shareholders attach greater weight
to the variable BVE for evaluation purposes than to the adaptation option value.14
In this scenario, the company chooses to adapt the line of business to focus on the
most profitable alternative activities which, according to Jensen and Ruback (1983),
has an ex-post effect on firm value.
Assuming an inverse relationship between the importance of BVE and the
financial health of the company, Barth et al. (1998) argue that the explanatory
power of the BVE is incremental as the company’s financial health deteriorates.
Again the main determinant of enterprise value is the results. The evaluation model
proposed by these authors also encompasses both the BVE and the results variables;
however, the theoretical support moves away from the OM, given the restrictive
assumptions of this model (the autoregressive process assumed for the variable
results; the dynamic of the variable other information; and the principle of risk
neutrality). The results are, however, consistent with the OM.
Thus, the market value of the company will be a function of the variable BVE
and results. The BVE provides useful information on the (extra) capacity of the
company’s debt, constituting a valuable instrument through which lenders can
control the conduct of managers and shareholders, as the first acts in the interests of
the latter and function accordingly as a proxy for the value of assets recognized in
the balance sheet. The results are assumed to be the determining variable for the
evaluation in a scenario of good financial health, being a proxy for the value of
assets not yet recognized, not reflected in the balance sheet. For example, Barth
et al. (1998) refer to investments, such as R&D, advertising, the technological
expertise of the company, human capital and customer portfolio, as intangible
assets, given that they count as costs in the year they are incurred, in respect of
GAAP. Thus, according OM and FOM, they recognize the accounting conservatism
effect in the variables BVE and results.
The authors’ theoretical foundation derives from the fact that to undertake an
investment project, and. assuming that they are rational in their decision making,
managers only undertake new investment projects if these are associated with
expectations of abnormal return, i.e. a positive NPV. On the acquisition date, assets
are recorded at historical cost, and so the surplus (cash flow) expected at the
acquisition date will be reflected in the variable “net income” in the following years,
14
Assuming that the Xt variable is the variable results, the authors show that for a given level of
BVE, the MVE is a convex function of the results. Analytically:
MVEt =BVEt1 ¼c1 BVEt1 =BVEt1 þ c2 Xt =BVEt1 þ e
, MVEt =BVEt1 ¼ c1 þ c2 ROEt þ e
(with ε*=ε/BVEt−1). If the ROE rate falls to unsatisfactory levels, (γ2) is irrelevant for assess-
ment (approaches the zero value), and (γ1) is irrelevant for valuation (approaches the unit value
—it does not achieves a unit value because, for example, assets are recorded at the historical cost
and not at fair value). The variable BVE is now the determinant variable for evaluation, because
in this context the adaptation option is valuable. Similar reasoning is obtained, if it is now setting
the level of the variable results (Xt).
and subsequently, as shown in OM and FOM, via the CSR property, in the BVE
variable. Thus, in periods of good performance, the main determinant of the value
of a company is based on the unrecognized assets. However, when the company’s
financial health deteriorates, and given that these intangible assets alone do not have
value but are dependent on the overall value of the company, its value tends
towards zero. Therefore investment in this type of asset is very risky, especially in
technology-based companies.
Damodaran (2001) mentions that a combination of this investment profile, i.e.
investment in intangible assets, in particular in companies in the start-up phase, which
favours a growth maximization strategy and an increased debt, results in an explosive
mix that significantly increases the probability of company default, because at this
stage the results tend to be negative. An increase in indebtedness encourages owners
to invest suboptimally and to take decisions that destroy value, which increases the
risk to the company and decreases the debt value (Myers 1977). By investing in
projects with high returns and high risk, investors are essentially betting against the
money lenders. If the investment proves profitable, shareholders collect a significant
portion of earnings; if the project fails, it is the creditors who bear the costs.
Inspired by the pioneering work of Black and Scholes (1973: 649), Berger et al.
(1996) argue that shareholders hold a put option (which in America may be
exercised at any time) on future cash flows of the company. If in a given period the
cash flows fall short of the expectations of shareholders, they can exercise the put
option and liquidate the company’s assets. Authors such as Barth et al. (1998) and
Berger et al. (1996) recognize that it is particularly difficult to determine the value
of this put option because the liquidation value of the assets varies over time
depending on economic conditions and the specificity of them. They propose that
the value of the put option acts as a proxy for the liquidation value of the assets—
the exit value— is defined as a function of the following variables:
Exit value ¼ 1:00ðcash and credit titlesÞ þ 0:72ðclients) þ 0:55ðinventoriesÞ

þ 0:54ðfixed assets)1:00ðcurrent liabilitiesÞ1:00ðlong term liabilitiesÞ
ð3:4Þ
According to Myers and Majd (1990) and Shleifer and Vishny (1992), the value of
liquidation depends on the type of assets held by the company. The more general the
higher the liquidation value, so the greater the likelihood of exercising the aban-
donment option, which is also very sensitive to the financial health of the company, as
measured by the Z-score proposed by Altman (1968). It is in this context that Berger
et al. (1996), like Barth et al. (1998), postulated the power of the variable BVE as a
proxy for the liquidation value, when the financial health of the company degrades.
In this line of research, the results of Schnusenberg and Skantz (1998) are
interesting. These authors also adopt the OM, documenting an increase in the
explanatory power of the BVE not only for companies under financial stress, such
as those reporting losses for ten consecutive years but not declaring bankruptcy and
which are not settled, but for companies that, despite posting a good performance
with successive profits, opt for voluntary liquidation. In this context, the BVE is
particularly relevant for evaluation, because, to liquidate the company, shareholders
voluntarily sell to third parties the assets held by paying their debt (if any) and
dividing the remainder among the shareholders. Abnormal returns recorded by
these companies would reflect that the market had not anticipated this decision by
shareholders. For example, the authors suggest that voluntary liquidation by
shareholders is a response to a threat takeover.
The literature, in particular Francis and Schipper (1999), whose period of
analysis extends from 1952 to 1994, has documented a fall in the explanatory
power of the variable “net income” compared to the variable BVE, not ascribing to
the OM a decrease in their explanatory power. According to Collins et al. (1997),
the increase in the explanatory power of the variable BVE against the variable
results based on a sample of North American companies for the period 1953–1993
is based on four key factors: (i) an increase in the number of companies reporting a
higher volume of transitory items (annual item 17 in the Compustat database) and/or
extraordinary items (for example, the authors mention costs resulting from
restructuring and adapting business lines); (ii) an increase in the volume of
investments in intangible assets (e.g. R&D and advertising); (iii) accordingly, and
in compliance with GAAP, a simultaneous increase in the number of companies
reporting losses (conservatism accounting); and iv) these effects occur in a context
identified by Hayn (1995) which is characterised by a change in the business profile
operating in the market, i.e. small companies, mostly technology-based, whose
probability of default increases sharply, given that they report a higher volume of
losses and for longer periods. These results seem to reinforce the explanatory power
of BVE as a proxy for the liquidation value of the company.
However, given the results that were already obtained by Collins et al. (1999)
when assessing the explanatory power of the variable “BVE” as a proxy for future
results, and according to the OM, the author defined FTUX = r*BVEt−1 as a proxy
for liquidation value,15 which is calculated based on the liquidation value (exit
value) proposed by Berger et al. (1996) (Eq. 3.4). Note that there is not clear
predominance in terms of explanatory power of the variable BVE as a proxy for the
liquidation value over the variable “FTUX”, even in companies that report suc-
cessive losses. Similar results are reported by Tan (2004) when analyzing the
incremental explanatory power of the variables BVE and results in the context of
the OM model in two groups of companies in a financial stress situation: companies
that went bankrupt, and companies that have opted for a process of merger and
acquisition (M&A). For Tan (2004), and as suggested by Burgstahler and Dichev
(1997), the M&A process is an adaptation variable for companies with poor
financial performance, i.e. those recording successive losses.
15
When calculating the variable FTUX as a proxy for future earnings, Collins et al. (1999), to
estimate the cost of capital, considered the value of 8.67% as a risk premium, similar to Berger
et al. (1996).
Myers (1977) suggests that the report of losses and persistence of them are
associated with high growth opportunities and/or a high proportion of investment in
intangible assets. For Berger et al. (1996), their liquidation value (exit value) will be
very low due to their specificity, thus the value of intangible assets will be only
preserved just as long as the company remains in business as a whole.16 In this
context, and as demonstrated by Tan (2004), the M&A strategy preserves the value
of the company, which is reflected in increased wealth for the shareholders. Thus, to
those companies, the variable results show a strong association with price. As for
the information content of the variable BVE, and in line with the OM, it seems to be
more relevant as a proxy for future expected results and more significant for the
sample of firms that have opted for M&A processes.17 Tan (2004) concludes that
the BVE does not reveal a significant explanatory power in relation to the group of
companies making losses, which in part contradicts the results of earlier investi-
gations (Collins et al. 1997; Barth et al. 1998; Francis and Schipper 1999). The
explanation advanced by Tan (2004) is based on the restrictive nature of the sample:
companies that have gone bankrupt or have opted for a process of M&A.
Barth et al. (1998) noted that technology-based companies invest heavily in
intangible assets, thus it is not a clear border between the recognized assets mea-
sured by BVE and the unrecognized assets whose proxy is the variable results. As a
result of conservative accounting, these companies tend to have negative values for
the variable BVE. For the sample that went bankrupt, the authors found that 20% of
companies registered negative values for the variable BVE in the year before
bankruptcy. Given these results, they question the adequacy of the BVE variable as
a proxy for the liquidation value.
By analyzing the properties of the estimated coefficients for the variables BVE
and net income in the OM, Penman (1998) found that in 12 subgroups, the value of
the median variable BVE assumed negative values, results that were difficult to
interpret. These results are similar to Amir and Lev (1996), Francis and Schipper
(1999) and Frazen and Radhakrishnan (2009) who for some years also reported
negative coefficients for the variable BVE. Zhang (2000) shows analytically that
these results are a consequence of strict accounting conservatism, particularly in
technology-based companies in the start-up phase, reflecting the necessary adjust-
ments in terms of additional investments in operating assets to ensure a flow of
positive liquid results in the future.
The sharp increase in companies reporting losses, particularly following 1990s,
was widely reported (e.g. Hayn 1995; Collins et al. 1997, 1999; Frazen and
Radhakrishnan 2009; Chan et al. 2001; Joos and Plesko 2005, among others). In
order to analyze the performance of the OM model for this type of company,
16
Ramsey and Shapiro (2001) provide empirical evidence of this effect in the aerospace industry.
Given the specificity of its assets, these tend to be sold at a low value in the event of liquidation.
17
Tan (2004) to determine the likelihood of M&A resorted to proxy defined by Pestana and Ruland
(1986), in which the probability of M&A versus failure is defined according to the percentage of
shares held by the managers, the level of debt, the size, and the magnitude of the tax losses that
may be deducted in future years.
Table 3.2 McCallig (2004) criteria for splitting the sample

Current results
Retained earnings Profits Losses
Positive Maturity firms (I) Firms candidates to exit the market (III)
Losses Growth firms (II) Revenue investment firms (IV)
McCallig (2004) begin by splitting the sample in terms of (1) report of profits or
losses and (2) the nature of retained earnings (positive or negative). Thus four
business groups were identified: Group I: profitable companies with positive
retained earnings, which ranks as companies at maturity; Group II: companies
which have positive results, but with accumulated losses, indicating growing
businesses; Group III: companies reporting losses and positive retained earnings,
which shows a deterioration in economic conditions, making these companies
candidates for exit the market; and Group IV: companies that report losses and also
have negative retained earnings, called revenue investment firms. This situation is
summarized in Table 3.2.18
Focusing the analysis on the latter group of companies (group IV), McCallig
(2004) postulate that the losses are a consequence of a strategy of aggressive
investment in R&D. Given the persistence of this profile of investment in intangible
assets, and in agreement with Zhang (2000), the variables results and BVE are
subject to the conservatism accounting effect, given the cumulative effects over time
of losses. In this context, the BVE, and as suggested by the OM, is not an adequate
proxy for future results.
Empirically, and with reference to the OM, McCallig (2004) disaggregates the
variable BVE in order to isolate the retained results (negative), which reflect the
intangible assets investments made in the past. The empirical evidence obtained
confirms the effect of positive valuation of losses (either for the variables “results”
or “negative retained results”). As for the variable BVE, when they isolated the
effect of retained earnings now as expected and according to the OM or the FOM
reports a positive and statistically significant coefficient. McCallig (2004) further
cites as an example the investigation by Burgstahler and Dichev (1997), in which
the nonlinear relationship between the variables MVE and BVE is clearly due to the
conservatism accounting to which the variable BVE is subject, in particularly in
technology-based companies in start-up or growth phase, which he calls revenue
investment firms. McCallig (2004) concludes that this group of companies are
younger companies (with an average age of seven years), small, with greater
specificity of assets, low fixed assets and with high growth opportunities, judging
by the large investments in R&D and advertising. The author even suggests that
18
Similar to our research (see Sect. 6.3), the variable “results” is a proxy for the stage of the firm’s
life cycle.
while in the 1970s companies tended to register small amounts of losses and for
short periods of time, after the 1990s, the losses tended to last for longer periods,
assuming greater magnitude, which, in the opinion of McCallig (2004), justifies the
persistence of the phenomenon of positive assessment of losses by the market. With
this research, McCallig (2004) recognized the phenomenon of positive valuation of
losses by the market so far identified as an abnormal phenomenon, despite several
investigations having already detected this phenomenon, and not only in new
economy companies.
Collins et al. (1999) referring to the research by Hayn (1995) demonstrated that
the effect of positive evaluation of losses detected by Hayn is due to an incorrect
specification of the earnings capitalization model used. Citing Greene (2000),
Collins et al. (1999) argue that when a relevant variable is omitted from the model,
and the variable is positively correlated with the dependent variable (MVE) and
negatively with the independent variable included in the model (results), the
exclusion of this variable from the model induces a negative bias of the estimated
coefficient of the independent variable—results. In this context, it is relevant to
include the variable “BVE” in the valuation models. They maintain that the effect of
positive valuation of losses is due to the omission of the variable BVE in the
evaluation model. In line with McCallig (2004), Joos and Plesko (2005) designate
positive evaluation of losses as a new phenomenon. In a replication of the study by
Hayn (1995), the authors documented a favourable price reaction to losses which
tends to revert quickly in profitable companies (in line with the contemporary
literature), but the price also reacts favourably for companies that report successive
losses, which contradicts the theory of the abandonment option. The authors
demonstrate that for systematic losses recorded for the second sample, for the
period 1990–2000, investors tend not to evaluate the variable “net income” on
aggregate, but rather its various constituents, in particular R&D, which the authors
identify as a proxy for future growth opportunities. Joos and Plesko (2005) con-
clude therefore that the persistence of the losses is associated with massive
investments in R&D and advertising, especially after the 1990s. In the 1970–1990
sample, the percentage of companies reporting losses was 15%, a value that is
duplicated for the period 1990–2000. This persistence of the losses means that they
cease to be an appropriate proxy for the highest probability of liquidation, even after
controlling for extreme observations via rank regressions.
Table 3.2 summarizes the current research which sees the losses, or rather their
persistence, as an appropriate proxy for the financial liquidation option held by
shareholders. In this context, the BVE is relevant as a proxy for liquidation value. It
is important to stress that these investigations tend to focus their analysis on the
period prior to 1996, which in our research is identified with the beginning of the
NEP. We also include investigations whose sample period covers the entire 1990s
in which losses tend to persist for longer periods. For this group of companies—the
new economy companies (the subject of study), usually younger companies, small,
technology-based, with greater specificity of assets and significant investments in
R&D, the effect of positive valuation of losses is statistically significant, i.e. the
higher the volume of losses the greater is the market capitalization of the company.
Table 3.3 Summary of the various studies on the information content of the variables earnings
and book value of equity for evaluation purposes
Current losses, they tend to be transitory Given the increased number of companies reporting losses
and so irrelevant to evaluation. If they (particularly after the 1990s), which are explained in part by
persist, shareholders exercise the liquidation high investments in items such as R&D and advertising,
option they hold on the company’s assets. recorded as costs in the year they are incurred, in line with
In this context, the variable BVE is a proxy GAAP, losses cease to be a consistent proxy for the financial
for the financial liquidation option of the liquidation option. The phenomenon of positive valuation of
company. The phenomenon of positive losses tends to emerge in association with technology-based
valuation of losses may occur, seen as an companies in the start-up phase, a result of the effect of
anomaly accounting conservatism. So for assessment purposes, it is
relevant to disaggregate the variable ‘net income’ into its
constituents
BVE = Proxy for the liquidation value of

the company
BVE (?) = Proxy for the liquidation value of the company

Authors Model(s) Period Authors Model(s) Period
Non-net firmsa Non-net firmsa
Hayn (1995) Returns 1963–1990 Franzen and OM 1980–1998
Radhakrishnan
(2009)h
Subramanyam Returns 1981–1990 Collins et al. OM 1975–1992
and Wild (1999)i
(1996)
Chambers Returns 1976–1992 McCallig (2004) OM 1980–1997
(1997)b
Basu (1997)c Reverse 1963–1990 Core et al. Model 1975–1999
regression (2003)j
Berger et al. Model 1984–1990 Tan (2004)k OM 1985–1997
(1996)d
Collins et al. OM 1953–1993 Joos and Plesko OM returns First sample
(1997) (2005) 1970–1990
Burgstahler OM 1976–1994 Second sample
and Dichev 1990–2000
(1997)
Barth et al. OM 1974–1993 Net firmsl
(1998)e
Schnusenberg OM 1975–1994 Trueman et al. OM 1998Q1–2000Q2
and Skantz (2000)
(1998)
Collins et al. OM 1975–1992 Hand (2001)m OM 1997Q1–2000Q3
(1999)
Francis and OM 1952–1994 Demers and Lev Model Three dates
Schipper (2001)n 03/12/1999
(1999) 28/2/2000
31/05/2000
(continued)
Table 3.3 (continued)

Amir and Lev OM and 1988–1993 Knauff and Goot Model 1998–2000
(1996)f others (2001)o
Kothari and Price and 1952–1989 Kozberg (2009) OM 02/1999–05/2001
Zimmerman return
(1995)g models
Tockic (2004, Model from Miller 1996–2000
2005) and Modigliani
(1961)
Martínez and OM 1996Q1–2001Q2
Clemente (2002)
Hand (2003) OM (adapted) 1997Q1–2000Q3
a
The designation “non-net firms” here is not restricted to the concept of the control sample used by us (see
Sect. 5.3.2)
b
Note that Chambers (1997) excludes from the analysis companies in the start-up phase
c
The objective of Basu (1997) is to assess the impact of accounting conservatism in β (PER). The conclusion
reached shows that losses (effective and potential), if any, are transitory, restricting their effect to one year in
accordance with the principle of prudence
d
The authors adopt multiple linear regression models, which explain whether the MVE on the basis of variables
such as the present value of the expected cash flows, the exit value, excess book value = BVE—exit value,
among others. They also explain the value of the abandonment option depending on the type of asset. We did
not explain here the exhaustive list of all explanatory variables used
e
The authors adopt the OM, but under a different theoretical framework, to avoid restrictive assumptions
underlying the OM (e.g. the autoregressive process for the abnormal results and neutrality towards risk). They
obtained conclusions in line with the OM
f
These authors detect the effect of positive valuation losses (see their Table 2). They justify this result due the
emerging nature of the sector at the time and rapidly growing—mobile phone companies. It is one of the first
investigations to suggest the introduction into the valuation models of non-financial variables, for example the
percentage of the population already covered by the network of mobile phones or the number of customers
g
These authors have detected the effect of positive valuation of losses (see their Table 5), but do not provide any
explanation for it
h
Frazen and Radhakrishnan (2009) assess the explanatory power of the OM applied to four groups of
companies that report losses: (1) companies that reported losses in a single year; (2) companies that reported
successive losses and are involved in a process of recovery/adaptation; (3) companies in liquidation/bankruptcy;
and (4) investment losses firms which, as in McCallig (2004), are companies that invest heavily in intangible
assets, while remaining still in a start-up/growth phase
i
These authors do not demonstrate a clear predominance in terms of the explanatory power of the exit variable,
calculated as proposed by Berger et al. (1996), as a proxy for liquidation value compared to the variable
FUTX = r * BVEt-1, as suggested by the OM, or as a proxy for future earnings, even in companies in financial
stress (i.e. reporting successive losses) which are candidates to exit the market
j
The authors do not adopt a specific theoretical model (e.g. the OM); however, they do detect the phenomenon
of positive valuation losses: when introduced into the model as an independent variable, the variable results and
it assumes negative values. The authors explicitly state that for 1999 the sample with losses includes 214
companies of the new economy, meeting the selection criteria adopted by Hand (2001)
k
Tan (2004) focuses his research only on companies that have gone bankrupt or have opted for a process of
M&A
l
They list (but without limitation) investigations that focus on net firms, which are our object of analysis. It
should be noted that, given the emerging nature of this sector, data were obtained on a quarterly basis
m
The author adopts the model of OM and according to Amir and Lev (1996) includes also non-financial
variables: web traffic variables, i.e. number of visitors per page, number of hours spent on the given site,
number of pages visited and the percentage of Internet user population
n
The authors did not adopt any formal evaluation model. A couple of web traffic variables (non-financial) are
selected based on factor analysis, checking the effect of positive valuation of losses on the variable operating
cash flow
o
The particularity of this research is that it deals with the IPOs of European companies listed in the index
EURO.NM. In order to explain the ratio MVE/BVE at the IPO date, the authors find that, for the variables
results, operating cash flow and earnings before extraordinary items taken in isolation to avoid the effects of
multicollinearity, the effect of positive evaluation of losses is reported
The theoretical support for this, as demonstrated by the OM and FOM, results from
the accounting conservatism effect, in which losses are due to high investments in
intangible assets (R&D and advertising), recorded as costs in full in the year in
which they are incurred, in obedience to GAAP, but which for the market signal the
probability of existence of higher growth opportunities (for the impact of the
expected future abnormal results (FE) in the ratios P/B and PER, see Fig. 2.1).
In the next chapter, given the higher magnitude that investments in R&D and
advertising assume, which are, according to Richardson and Tinaikar (2004), the
main cause of the conservatism accounting modelled by OM and FOM, especially
in technology-based companies in the start-up phase, we start by analyzing the
impact of these investments on the market value of the company. In Sect. 4.3, we
evaluate the effect of growth on the market value of the company, with a particular
focus on the internet sector, highlighting the growth potential of this sector which is
associated with massive investments in the intangible assets R&D and advertising.
Table 3.3 presents a summary of the relevant studies on the information content of
the variables earnings and BVE for evaluation purposes.
References
Altman EI (1968) Financial ratios, discriminant analysis and the prediction of corporate
bankruptcy. J Financ 23(4):589–609
industry. J Acc Econ 22(1–3):3–30
Ball R, Brown P (1968) An empirical evaluation of accounting income numbers. J Acc Res 6
(2):159–1478
Ball R, Watts RL (1972) Some time properties of accounting income. J Financ 27:663–682
Barth M, Kallapur S (1996) The effect of cross-sectional scale differences on regression results in
empirical accounting research. Contemp Acc Rev 13:527–567
Barth M, Beaver W, Landsman W (1998) Relative valuation roles of equity book value and net
income as a function of financial health. J Acc Econ 25:1–34
Basu S (1997) The conservatism principle and the asymmetric timeliness of earnings. J Acc Econ
24(1):3–37
Beaver WH (1968) Market prices, financial ratios and the prediction of failure. J Acc Res 6:179–
192
Beaver WH (2002) Perspectives on recent capital research. Acc Rev 77(2):453–474
Beaver WH, Lambert H, Morse D (1980) The information content of security prices. J Acc Econ
2:3–28
Berger PG, Ofek E, Swary I (1996) Investor valuation of the abandonment option. J Financ Econ
42:257–287
Black F, Scholes M (1973) The pricing of options and corporate liabilities. J Polit Econ 181:637–
654
Burgstahler DC, Dichev I (1997) Earnings, adaptation and equity value. Acc Rev 72(2):187–215
Chambers DJ (1997) The information content of negative earnings and its relation with initial-loss
persistence. Working Paper, University of Illinois at Urbana-Champaign. http://ssrn.com/
abstract=15048
Collins D, Kothari SP (1989) An analysis of the intertemporal and cross-sectional determinants of

the earnings response coefficients. J Acc Econ 11:143–181
Collins D, Kothari SP, Shanken J, Sloan RG (1994) Lack of timeliness versus noise as
explanations for low contemporaneous return-earnings association. J Acc Econ 18:289–324
Collins D, Maydew EL, Weiss IS (1997) Changes in the value-relevance of earnings and book
values over the past forty years. J Acc Econ 24(1):39–67
Collins D, Pincus M, Xie H (1999) Equity valuation and negative earnings: the role of book value
Core JE, Guay WR, Buskirk AV (2003) Market valuations in the new economy: an investigation
of what has changed. J Acc Econ 34(1–3):43–67
Damodaran A (2001) The dark side of valuation. Prentice Hall, United States
Demers E, Lev B (2001) A rude awakening: the internet shakeout in 2000. Rev Acc Stud 60(2–
3):331–359
Donnelly R (2002) Earnings persistence, losses and the estimation of earnings response
coefficients. Abacus 38(1):121–133
Easton P, Harris TS, Ohlson J (1992) Accounting earnings can explain most of security returns: the
case of long event windows. J Acc Econ 15:119–142
Ferreira L, Sarmento M (2004) Que modelos adoptam os Revisores Oficiais de Contas na
Avaliação de Empresas e de Participações Sociais? Revisores e Empresas 6(25):19–31
Francis J, Schipper K (1999) Have financial statements lost their relevance? J Acc Res 37(2):319–
352
Franzen L, Radhakrishnan S (2009) The value relevance of R&D across profit and loss firms.
J Acc Public Policy 28:16–32
Greene WH (2000) Econometric analysis, 4th edn. Prentice Hall, New York
Guimarães JFC (2004) A contabilidade: utilidade para a gestão (decisão). Revisores e Empresas
25:44–50
US internet stocks. Eur Financ Rev 5(3):295–317
Hand JRM (2003) Profit, losses and non-linear pricing of internet stocks. In: Hand JRM, Lev B
(eds) Intangibles: an Oxford management reader. Oxford University Press, Oxford
Hayn C (1995) The information content of losses. J Acc Econ 20(2):125–153
Jensen M, Meckling W (1976) Theory of the firm: managerial behaviour, agency costs and
ownership structure. J Financ Econ 3(4):305–360
Jensen M, Ruback R (1983) The market for corporate control: the scientific evidence. J Financ
Econ 11:5–50
Joos P, Plesko GA (2005) Valuing loss firms. Acc Rev 80(3):847–870
Knauff P, van der Goot T (2001) The relevance of reported financial in-formation for valuing
European IPOs. EFMA Lugano Meetings. http://ssrn.com/abstract=263176
Kormendi R, LipE RC (1987) Earnings innovations, earnings persistence and stock returns. J Bus
60:323–345
Kothari SP (2001) Capital markets research in accounting. J of Acc Econ 31(1–3):105–231
Kothari SP, Zimmerman JL (1995) Price and return models. J Acc Econ 20:155–192
accounting and non-financial information in explaining revenues and valuations for internet
firms. VDM
Martínez FG, Clemente M (2002) The added value of non-financial information in internet firms
pricing. Working Paper, Polytechnic University of Valência Department of Economics and
Social Sciences. http://ssrn.com/abstract=328781
McCallig J (2004) Revenue investment, accounting conservatism and the valuation of loss making
firms. Paper presented at 14th annual conference on financial economics and accounting
(FEA). http://ssrn.com/abstract=488105
References 63
Modigliani F, Miller M (1966) Some estimates of the cost of capital to the electric utility industry
1954–1957. Am Econ Rev 56(3):333–391
Miller, M. and Modigliani, F. (1961): ”Dividend Policy, Growth and the Valuation of Shares”,
JBus, 34(4), 411–433
Myers S, Majd S (1990) Abandonment value and project life. Adv Futures Options Res 4:1–21
Neves JC (2002) Avaliação de empresas e negócios. McGraw-Hill, New York
(2):661–687
Ohlson J, Penman S (1992) Disaggregated accounting data as explanatory variables for returns.
J Acc Auditing and Financ 7:553–573
Ohlson J, Shroff PK (1992) Changes versus levels in earnings as explanatory variables returns:
some theoretical considerations. J Acc Res 30:210–226
Penman SH (1998) Combining earnings and book value in equity valuation. Contemp Acc Res
15:291–324
Pestana V, Ruland W (1986) The merger/bankruptcy choice. Acc Rev 61:288–301
Ramakrishnan R, Thomas JK (1993) Valuation of permanent, transitory and price-irrelevant
components of reported earnings. Working Paper. Columbia University, New York
Ramsey VA, Shapiro DM (2001) Displaced capital: a study of aero-space plant closing. J Polit
Econ 109:110–118
Financ 44(2):223–333
Schnusenberg O, Skantz TR (1998) A test of the abandonment hypothesis using voluntary
liquidating firms and unprofitable surviving firms. J Acc Auditing Financ 13(4):395–415
Shleifer A, Vishny RW (1992) Liquidation values and debt capacity: a market equilibrium
approach. J Financ 47:1343–1366
Subramanyam KR, Wild JJ (1996) Going-concern status, earnings persistence, and informative-
ness of earnings. Contemp Acc Res 13(1):251–273
Tan C (2004) Alternative adaptation of distressed firm’s resources: the valuation roles of equity
and earnings. Working Paper, New York University. http://ssrn.com/abstract=293125
Tokic D (2004) R&D, advertising and the market value of internet firms: Part I. J Internet Commer
3(2):79–99. Doi:10.1300/J179v03n02_05
Tokic D (2005) R&D, advertising and the market value of internet firms: Part II. J Internet
Commer 4(4):23–40. doi:10.1300/J179v04n04_02
Trueman B, Wong MHF, Zhang X (2000) The eyeballs have it: searching for the value in internet
stocks. J Acc Res 38(Supplement):137–162
Wiley, New York
Wysocki PD (1998) Real options and the informativeness of segment disclosures. Working Paper.
University of Michigan, MIT
Chapter 4
The Impact of Investment in Intangible
Assets on the Market Value
Abstract Given the higher magnitude of investments in R&D and advertising that
are assumed in net and non-net firms, this chapter analyses the impact of these
investments on the market value of a company. We also evaluate the effect of
growth on the market value of the company, with a particular focus on the Internet
sector, highlighting the growth potential of this sector due its emerging nature.
Keywords Investments in R&D R&D and market value Growth Internet as

an emerging industry
4.1 Introduction
As presented in the Chap. 2, the FOM demonstrated analytically that goodwill, as

measured by the difference between the MVE and the BVE is the result of three
factors: (i) persistence of abnormal future results; (ii) other non-financial infor-
mation that is coming to the market, which is incorporated immediately into prices
and is later recognized in the financial statements (lack of timeliness—variable vt);
and (iii) the underestimation of the operating assets (accounting conservatism).
With reference to non-operating assets, given the existence of liquid markets, the
book value reflects the firm’s market value.
The effect of the undervaluation of assets becomes relevant with the development
of technology-based companies (high-tech firms). In this type of company, invest-
ments tend to be concentrated in R&D, marketing and advertising in order to promote
the image of the company/services/products. In the absence of objective criteria that
could be used to measure the future benefits associated with this type of investment (if
any), GAAP accounting principles require their treatment as costs in full in the year
they are incurred. These procedures tend to underestimate the net profit, the value of
assets, and the value of equity, particularly in companies in the growth phase, and
which causes a strong distortion in the information provided by financial statements.
In Chap. 3, our analysis focused on the information content for evaluation pur-
poses of the variables net income and BVE as major determinants of value according

DOI 10.1007/978-981-10-3009-3_4
66 4 The Impact of Investment in Intangible Assets on the Market Value
to OM and FOM, with particular emphasis on the group of companies reporting

losses. Based on the literature review, it appears that the phenomenon of positive
valuation of losses is recent (have assumed relevance only in the 1990s), and is
associated with start-up, technology-based companies with high growth opportu-
nities. For growing businesses, it is important to disaggregate the net income into its
constituents, in particular R&D and advertising, which for accounting purposes are
considered costs, but which the market tends to evaluate as assets.
The value of intangible assets, including investment in R&D, advertising,
intellectual capital, strategic partnerships, customer base established contracts,
goodwill, location, brands and patents, etc. becomes decisive for the companies to
innovate, differentiate products and services, and develop customer loyalty: in
short, to create value. Thus, given the importance of this type of investment, this
chapter begins by reviewing the impact of such investment on the MVE. Next, and
given the emerging nature of the Internet sector/net firms, we focus attention on the
growth potential for evaluation purposes. According to Copeland et al. (2000),
Damodaran (2001) and Jorion and Talmor (2006), among others, at this stage,
companies in general (not just the new economy companies) tend to adopt a growth
maximizing strategy. Given their profile, technology-based companies’ growth
appears to be associated with massive investments in intangible assets, such as
R&D and advertising.
4.2 The Impact of Investment in Intangible Assets

on the Market Value
The value of a company should reflect the value of its net assets. A modern
economy, according to Lev and Sougiannis (1996) and Chan et al. (2001), is
characterized by the emergence of new sectors: sectors based on knowledge,
software, and those related to biotechnology and telecommunications; therefore the
value of these companies, commonly known as technology-based companies or
high-tech firms, is mainly associated with the intangible assets they hold. As
intangible assets, these authors refer to R&D and promotional campaigns and
marketing (advertising) in the traditional studies on the industrial economy.
The prevalence of such assets makes the task of evaluation of this type of
company particularly complex, since the recognition of an intangible asset is not a
straightforward question: it has to be identified, even though it is without physical
substance. In this sense, the costs of R&D cannot be associated with intangible
assets, as they do not necessarily exist. It turns out that the research phase is often
long and fruitless. According to FASB, this high degree of uncertainty about future
economic benefits justifies why this kind of investment is not capitalized.1
1
It is not the objective of this study to analyze the suitability or not of the GAAP currently in the
US. The main arguments to support the accounting as expenses (rather than capitalization) of
4.2 The Impact of Investment in Intangible Assets on the Market Value 67
Given these accounting practices, White et al. (1997), Copeland et al. (2000),
Damodaran (2001), and Chan et al. (2001) among others argue that the traditional
evaluation metrics, such as the multiple of results (PER) and the multiple of book
value (P/B), suffer great distortion and volatility given the associated strong-growth
expectations associated with those companies.2 Chan et al. (2001) document for the
period between 1975 and 1995, and with reference to the American companies listed
on the NYSE (New York Stock Exchange), AMEX (American Stock Exchange) and
NASDAQ (National Association of Security Dealers’ Active Quotation), that the
ratios of R&D over sales and R&D over the BVE, ratios that measure the intensity of
investments in intangible assets, show an increased from 1.7 and 4.13 in 1975 to 3.75
and 10.88 in 1995, respectively. This effect seems to be higher for certain sectors, i.e.
737—computer programming, software and services; 283—drugs and pharmaceu-
ticals; 357—computers and office equipment; 38—measuring instruments; 36—
electrical equipment, excluding computers; 48—communications and 37—trans-
portation equipment. Lev and Sougiannis (1996) also observe high growth in R&D
and advertising expenses, particularly over the last 20 years, as a result of the
emergence of new sectors. Similar empirical evidence is provided by Damodaran
(2001: 9), whose comparative analysis includes the Internet sector. Thus, given the
magnitude of technology-based companies in the current economy, it is important to
analyze the impact of the investment in intangible assets such as R&D and adver-
tising on the MVE of these companies.
In an efficient market, one would not expect any relationship between the
variables MVE, R&D and advertising, as prices tend to reflect all publicly available
information. However, markets are characterised by a set of imperfections, due to
the effect of taxes, transaction costs, asymmetry of information and agency costs
(Quintar and Zisswiller 1994); so it is reasonable to analyze the statistical rela-
tionship between these variables.
In this context, event studies seek to evaluate the market reaction to the
announcement of new investment projects in R&D by companies. The research by
Wooldridge (1988) was a pioneer study in this area. The author reported a positive
reaction from the market to new investment projects in R&D. Adopting a similar
methodology, Chan et al. (1990) and Zantout and Tsetsekos (1994) showed positive
and statistically significant returns for the group of high-tech firms contrary to the
results obtained for the group of low-tech firms. To guarantee robustness of the results,
intangible R&D and advertising are: (i) information asymmetry between managers and
shareholders and the resulting agency costs that exist not only between managers and shareholders,
but also between them and the lenders; (ii) contract costs between the parties, given once again the
privileged position of managers’ access to internal company information; (iii) fiscal motivations;
(iv) restrict the possibility of a discretionary management of results by managers (creative
accounting); and (iv) increasing the liability of auditors, given recent accounting scandals (Basu
1997; Kothari et al. 2002; Hand 2003).
2
Figure 2.1, shows how the ratios PER and P/B could achieve high values in this type of firm,
based on the expectation of high growth opportunities.
Szewczyk et al. (1996) used the Tobin’s Q ratio (which allows them to control the
industry effect) to analyze market reaction to the announcement of new R&D
investment projects.3 The results confirm the results recorded in 1994, i.e. the market
positively evaluates the investments in R&D and advertising in technology-based
companies. In the opinion of the authors, these results reflect perceptions by investors
of a higher probability of their existence in the portfolio of this type of business with
new growth opportunities, confirming the investment opportunity hypothesis. The
negative returns or a decrease in the value of equity registered for the group of
low-tech companies are justified by the authors as a consequence of agency conflicts
on the free-cash flow. As postulated by Jensen (1986), the announcement of new
investments in R&D and advertising in companies with high free-cash flow, or
companies in general at maturity, intensifies the agency conflicts between managers
and shareholders. The managers usually prefer to invest more even in low profitability
projects, or even accept projects with negative NPV because the greater the volume of
resources, the more power the management has. Investors prefer the distribution of the
excess liquidity, whether in the form of dividends and/or the repurchase of shares. In
addition to investigations based on the event studies, cross-sectional and/or pooled
regression studies also analyze the existence of a statistically significant relationships
between the MVE and the variables R&D and advertising.
Weiss (1969) argued that investments in R&D and advertising should be capi-
talized. But the empirical evidence in the 1970s showed inconclusive results. Peles
(1971) detects a positive statistical significance for investments in R&D for the beer
and tobacco sectors, revealing an insignificant statistical relation in the automobile
sector. Abdel-Khalik (1975) reports a statistical significance for the variable R&D
within the food industry and cosmetics, but no statistical significance for the
tobacco sector, thus contradicting the results of Peles (1971). Picconi (1977) does
not detect a statistical relationship between the variables R&D, sales and MVE.
Sougiannis (1994) justify this contradiction in terms of empirical results with ref-
erence to the 1970s due to the small size of the samples used and some compu-
tational limitations in terms of econometric treatment. From the mid-1980s, the
results seem to be more solid, confirming that the items R&D and advertising are
relevant in determining the value of companies.
Dukes (1976) analyzed the perceptions of investors to information provided by
the variables R&D and advertising and concluded that investors consider these
variables when they formulate their expectations about the value of businesses,
adjusting the net income reported by companies to the value of these variables.
Ben-Zion (1978) found that the difference between the MVE and the BVE (un-
recorded goodwill, according to the OM and FOM) is positively correlated with the
ratio of R&D over sales. Following this line of research, Hirschey (1982) argued that
the MVE depends on several signals sent to the market about the future profitability
prospects of the company. The items R&D and advertising are potential indicators of
3
In the study developed in 1994, Zantout and Tsetsekos used a dummy variable to differentiate the
two groups of companies, high-tech and low-tech firms.
the future abnormal returns. His study showed that the ratio MVE to book value (a
proxy for Tobin’s Q) is explained on the basis of the results obtained by the com-
pany, its volatility, the degree of industry concentration measured by the share of
sales of the four largest companies in the industry, the growth rate of sales, and the
variables R&D and advertising. As expected, the variables R&D and advertising are
associated with positive and statistically significant coefficients. The author con-
cludes that these variables are relevant to determine the value of the company.
In an extension of this work, co-authored with Connolly in 1984, the authors
maintain that the difference between the MVE and the book value (BV) of tangible
assets4 which they define as the surplus value (which we can identify with the
unrecorded goodwill defined by the OM and FOM) is explained by the variables
R&D and advertising and by the level of industry concentration. Analytically:
ab Surplus Value (unrecorded goodwillÞ

¼ f ðR&D; Advertising, Industry ConcentrationÞ
cd Sales
ð4:1Þ
The statistical significance associated with positive coefficients for the estimated
variables validates the results obtained in 1982. The authors interpret this result, as
anticipation of the NPV associated for this type of investment by the market. This
effect is strong if the company has a dominant position in the sector (Connolly and
Hirschey 1984). This dominance, coupled with successive investments in intangible
assets, enables the company to create barriers to entry and so perpetuate its com-
petitive advantages, thus ensuring the capacity to obtain economic rents.
In order to introduce the industry effect and the dimension effect into the analysis
on the assumption that the R&D variable is more effective in differentiating between
the industrial and the non-industrial firms, Chauvin and Hirschey (1993) propose a
model in which the ratio of MVE over sales is explained by the cash flows generated
by the company, the sales growth, the risk, the market share of the company, and the
variables R&D and advertising. The model is evaluated for the two groups of
companies, industrial companies, where the variable R&D becomes relevant and
non-industrial companies, for which advertising investments tend to dominate. The
model also considers the effect of size, breaking down the sample into large, medium
and small businesses. The results point to a strong statistical significance for the
variables R&D and advertising, whether in industrial or non-industrial companies.
The results are robust, confirming a positive statistical significance, regardless of
company size. The authors conclude therefore that the variables R&D and adver-
tising are useful to investors in formulating their expectations about the magnitude of
future cash flows and the level of risk associated with them. Table 4.1 systematizes
the main results obtained by Hirschey (1982, Connolly and Hirschey 1984, Chauvin
and Hirschey 1993).
4
The authors assume, however, that the book value of tangible assets is an imperfect proxy for the
market value thereof.
Table 4.1 Relationship between MVE and the investment in R&D and advertising
70
Author Hirschey (1982) Estimated Connolly and Hirschey (1984) (Estimated Chauvin and Estimated coefficient
Variable coefficient coefficient) Hirschey (1993)
Independent MVE/BV (MVE-BV)/sales MVE/sales
variable (unrecorded goodwill)
Dependent 1/BV 1/sales 1/sales
variables
Net income/BV Positiveg Cash flow/sales –Positive (L, M and S)g,e
–Positive (R&D, Adv)g,f
Level of Negative Level of concentration in the industry (GC)a Negativeg Share of the –Positive (L e M)
concentration in the market –Positive (R&D)g
industry (GC)a –Positive (Adv)g
Sales growth Positive Sales growth Positiveg Sales growth –Negative (L e M)
–Positive (R&D)g
–Positive (Adv)
Risk (volatility of Negativeg Risk (volatility of net income) Negativeg Riskd –Negative (L, M)
net income) –Negative (R&D e Pub)
R&D/BV Positiveg R&D/sales Positiveg R&D/sales –Positive (L, M and S)g
Advertising/BV Positiveg Advertising/sales Positiveg Advertising/sales
–Positive (L, M and S)g
Nonlinear effect for the variable level of Positiveg
concentration in the industry [(GC)2]b
Diversification
R&D/Sales * GCc Negativeg
c
Advertisement/sales * GC Positiveg
c
Sales growth * GC
a
The level of concentration in the industry is measured by the ratio of the sales of the largest firms in the industry over the total sales of the industry
b
The nonlinear effect is obtained by the square of the variable [(GC)2]
c
Interactive variables
d
The variable risk is measured by the methodology of Garman and Klass (1995), i.e. ln of the difference between the high price and low price for 52 weeks
e
L Large firms, M Medium firms, S Small firms
f
4 The Impact of Investment in Intangible Assets on the Market Value
R&D—identify the group of firms that invest more in R&D, i.e. industrial firms; Adv identify the group of firms that invest in advertising (non-industrial firms)
g
Variable is statistically significant
In short, at this point we conclude that, based on the results obtained by the event
study methodology, (e.g. Wooldridge 1988; Chan et al. 1990; Zantout and
Tsetsekos 1994 and Szewczyk et al. 1996) or based on the results from the
cross-sectional studies and/or pooled regression (Hirschey 1982; Connolly and
Hirschey 1984; Chauvin and Hirschey 1993), the market positively values a
company’s investment in R&D and advertising. This positive valuation reflects the
market expectations about the growth opportunities held by the company, and
which derive from the existence of licenses, patents, specific knowhow, innovation
and the capability of achieving significant market share, which require continued
investments in R&D and advertising.
In this context, the market seems to evaluate companies according to their
growth potential, so investors should pay a premium for the growth opportunities
held by the company. Given the emerging nature of the Internet sector/net firms, we
focus attention on their growth potential. We begin the analysis with the classical
models of Modigliani and Miller (1961) and Malkiel (1963), establishing an
analogy with the OM and FOM, in order to evaluate the impact of growth. We note
that the growth potential of net firms is associated with strong investments in the
intangible assets, such as R&D and advertising. The rationality of this investment
strategy, popularized with the concept of “winner takes all” (Noe and Parker 2005),
is based on the strong expectation of increasing returns to scale, enhanced by the
network effects generated by the Internet.
4.3 The Company Value and the Potential Growth
According to Modigliani and Miller (1961), the value of a growing company

(growth stock) is obtained by adding, to the present value of cash flows generated
from existing assets, the present value of growth opportunities held by the com-
pany. Analytically:

X k r bX
V0 ¼ þ ; ð4:2Þ
r r r
where
X net income generated by the company due to the assets in place;
k expected rate of return from new investments (k > r, given the presence of
investment opportunities);
r cost of equity;
b the percentage of retained earnings by the firm to finance new investments.
Assuming that the firm retains all of the net income generated (b = 1), a con-
dition that second Modigliani and Miller (1961) is a reasonable assumption given
the growth phase of the firm, we obtain:

X kr X Xk X
V0 ¼ þ ¼ ¼ m; ð4:3Þ
r r r rr r
where m = k/r, a multiple of profitability. As the company moves towards maturity,

growth opportunities tend to fade, so r = k, and the company’s value is the result of
the capitalization of the results generated by assets in place, equivalent to the
familiar Gordon model (a specific case of the OM) and consequently equivalent to
the RIV model, as demonstrated by Lo and Lys (2001).
When studying the sharp decrease in prices that occurred in 1962, Malkiel
(1963) demonstrates that the multiple of profitability (m) can be defined as:
ð1 þ gÞN
m¼m ; ð4:4Þ
ð1 þ rÞN
where
m the multiple of profitability of a company at maturity;
g growth rate;
r the cost of equity, assuming therefore g > r (given the existence of growth
opportunities);
N the number of periods during which it is expected to obtain abnormal returns.
Malkiel (1963) shows three properties associated with growth companies: (i) the
multiple of a growing company (m) is greater than a for company at maturity stage
ðmÞ whereby PER(m) > ðPER mÞ (as shown previously in Fig. 2.1 on the basis of
the RIV model); (ii) the multiple is a function of expected growth (g) and the
number of periods of expected abnormal growth (N) [PER = f (g, n)]; and (iii) the
volatility of (m) also depends on the growth rate (g) and the number of periods for
which abnormal growth is estimated (N) [volatility = f (g, N)]. The author con-
cludes that the price of such securities tends to be much more volatile [i.e. volatility
(m) > volatility ðmÞ].5
Comparing the Modigliani and Miller (1961) model with the OM, we conclude
that for both models the price of shares depend on the amount of existing assets
(BVE) and growth opportunities held by the company. However, the OM differs
from the Modigliani and Miller (1961) model, which assumes b = 1 (b—retained
earnings), so the growth rate is permanent. For the OM abnormal returns depends of
5
The motivation behind this study was to explain the sharp decrease of share prices in 1962. The
author concluded that the largest decrease in prices was reported by growth shares, a result of
strong volatility that characterised this type of shares. Ofek and Richardson (2002, 2003) reported
similar results for the group of net firms with reference to the crash that occurred in March, 2000.
4.3 The Company Value and the Potential Growth 73
an independent autoregressive process of the BVE variable incorporating shocks

from other non-financial sources of information via the variable vt, which tend to
converge quickly to the average profitability of the sector/industry due the com-
petition in the economy. Thus, since Mao (1966) and Taylor (1974), it has been
traditional to incorporate the concept of life cycle underlying the constant growth
assumption of the Modigliani and Miller (1961) model. Thus, in the first stage, the
company grows/invests at an exponential rate, while in the second stage the
company grows at a constant rate. In the final stage, in which growth opportunities
vanished, investments are restricted to the replacement of existing assets.
Comparing both models, both assume that the financing of new investment
projects by retained earnings and the issuance of new shares are perfect substitutes.
At maturity, the rate of return on equity (ROE) tends to equalize the cost of capital,
as evidenced by the Eq. 2.11, which defines the permanent results.
Feltham and Ohlson (1995) as an extension of the initial model of Ohlson
(1995), and in line with Malkiel (1963), demonstrate that the multiple of results in
growing firms is higher than the multiple of firms in maturity stage. This effect
occurs because: (i) the persistence of abnormal results, measured by the parameter
w11, which incorporates the growth effect (g) and duration of abnormal growth (N),
according to the terminology used by Malkiel (1963); (ii) the volume of invest-
ments made (the increase in operating assets), expressed by the parameter w22; and
(iii) the effect of conservative accounting, reflected in the parameter w12, a result of
the underestimation of operating assets. This effect is very pronounced in start-ups
firms particularly technology-based companies, given the predominance of intan-
gible assets.
The evaluation of a company should take into account its life-cycle stage (e.g.
Copeland et al. 2000; Damodaran 2001). Jorion and Talmor (2006) apply this view
to the new economy companies (firms under study) showing that in the
start-up/growth phase, these companies invest mainly in intangible assets, in par-
ticular in the items R&D and advertising. Therefore, given the increasing rise in
technology-based sectors, including the Internet, as documented by Lev and
Sougiannis (1996), Chan et al. (2001), Damodaran (2001), McCallig (2004) and
Joos and Plesko (2005), among others, and the fact that investments in this type of
company tend to concentrate the investments in intangible assets, and, given the
literature review presented in this section, companies that report losses cannot be
evaluated evenly.
The theoretical support for this reasoning derives from investments in intangible
assets being associated with a twofold effect, i.e. a short-term effect and an average
long-term effect. In the short term, investment in intangibles (e.g. R&D and
advertising) increases the company costs (given the GAAP requirement of imme-
diate accounting for them as costs, a procedure that tends to underestimate both the
company’s results and its assets). In the medium and long term, these investments
generate cash flows, reflected in the volume of sales, whose impact extends to the
results (Lev and Sougiannis 1996). Due this double effect, the literature suggests
that the different stages of the firm life cycle could be distinguished by the variable
results. Thus, profits are associated with companies that have reached maturity, or at
least steady growth, while losses are associated with younger companies in the
start-up phase, dominated by a growth maximization strategy.
McCallig (2004), given the magnitude of the reporting losses by US firms in the
1990s, uses the net income variable as a proxy to identify the company’s life-cycle
stage. The results show that for the group of companies reporting losses (and
accumulated losses), what he called revenue investment firms, the current and
accumulated losses are derived from the conservatism accounting, because this type
of company is dominated by the short-term effect associated with investments in
intangibles assets. Sougiannis (1994) and Lev and Sougiannis (1996) present
empirical evidence about the duality of short- and medium-term effects associated
with investments in intangible assets. Sougiannis (1994) showed that investment in
R&D is associated with profits, which reflect the benefits of the investments made
in the past. The results are conclusive, confirming that the market evaluates posi-
tively current investments in R&D, as a direct effect. The indirect effect results from
capitalization of the results generated by investments made in the past. The author
documents a stronger result for the indirect effect, finding that on average the
investment of a dollar in R&D provides an increase in net profits of more than two
US dollars within 2 to 7 years. Note that the greater expressiveness in statistical
terms of the indirect effect is a consequence of the fact that Sougiannis (1994) used
a sample of companies in the mature stage. Lev and Sougiannis (1996) point to a
time lag of 5 to 9 years for investments in R&D and advertising to start generating
positive cash flows.
However, despite the extensive literature documenting positive assessment by
the market of the items R&D and advertising (see Sect. 4.2, Table 4.1), Moore
(2002) suggests that such statistical relationships lack solid theoretical support.6
The FOM constitutes a theoretical framework that enables the explanation of the
unrecorded goodwill, i.e. the difference between the MVE and the BVE, which
reflects the present value of growth opportunities held by the company but not yet
recognized in the financial statements. Recall that following the FOM, the un-
recorded goodwill depends on the persistence of abnormal results and information
other than financial that come to market and is incorporated immediately into prices
and the conservative accounting effect (see Eq. 2.23; Richardson and Tinaikar
2004), which is persistent in technology-based companies in the start-up phase, as is
the case of net firms.
The MVE of the youngest companies does not reflect the effect of reputation, a
potentially valuable asset held by older companies (Diamond 1989). The imple-
mentation by the company of its obligations tends to signal to the market about the
growth prospects which reside in the variables R&D and advertising. The imple-
mentation of such investments by the company can be seen by the market as the
exercise of a call option (Moore 2002). Thus, this option can be defined as:
6
For example, Moore (2002) refers to the arbitrary method in the calculation of depreciation
adopted by Chan et al. (2001) in the value of assets or the amount of net investment in R&D and
advertising, when they propose to analyze the impact of the investment in R&D in the MVE of
technology-based companies in the start-up/growth phase.
PVGOðR&D;PubÞ ¼ Max½Et ðXÞC; O ð4:5Þ
where
– PVGO (R&D and advertising) is the present value of growth opportunities;
– C is the current investment to be made (the strike price);
– The underlying asset is identified with the investment project itself;
– Et(X) corresponds to the value of expected cash flows.
The estimation of the value of the variable Et(X) is complex. Thus, in the
absence of a criterion for measuring objectively the benefits associated with these
investments, this variable appears not to be reflected in the financial statements of
companies. However, assuming the principle of rationality, it is expected that
managers implement only in-the-money options, i.e. they invest in projects with
expectations of abnormal returns (with NPV positive), where it is expected that a
monetary unit invested generates an average rate of return greater than one invested
in equity (Graham and Harvey 2001).
The net firms’ strategy undertook massive investments in R&D and advertising.
For example, Hand (2003: 258) documents mean values for the ratio of R&D over
sales as ten per cent in companies with profits. In companies reporting losses, the
ratio amounted to 39%. For the aggregate ratios of sales over R&D and advertising,
the average is 37% for companies with profits, reaching 154% in companies with
losses. These data refer to the period 1997–2000 (Q3), including the post-crash
period of the “dot.com bubble”. Given these data, Trueman et al. (2000: 147) said:
As these statistics confirm, investors are clearly paying for growth rather than current
performance.
Investment in advertising aims to attract the attention of investors to enhance the

reputation and visibility of the company, and thus, increase web traffic in order to
increase future sales (e.g. online sales, revenue from services, provision of adver-
tising space on web pages, etc.). To measure the web traffic, multiple variables are
used. Trueman et al. (2000) used “number of unique visitors”, “number of pages
viewed”, and “time spent”. Hand (2001b) also used the variable “percentage of
population covered by Internet service”. The high interest of investors increases the
prices of shares, which enables the company to obtain additional funds to undertake
growth opportunities. Furthermore, the larger the size of the company, then the
greater is the confidence of investors in the company.
Investment in R&D is intended primarily for website design, software development,
and computer platforms in order to facilitate the navigation on websites, to ensure
safety, and simplify the online transaction process, and to create interactive software
(e.g. email alerts) in order to retain customers. The rationality of this investment
strategy derives from high-income expectations associated with the Internet network
effect, boosted by the fact that users are connected to a globally dimensioned network
R&D Advertising “Adv” by user R&D by user

(Adv)
(+) Audience (+) Time spend (>) N.º of visits by

by user web sites
(>) N.º the pages

view
(>) N.º the

advertisements
(>) N.º the

s
advertisement
view
(>) Sales
Adapted from Kozberg (2009).
Fig. 4.1 The effect of investment in R&D and advertising in abnormal future profitability based
on the networking effect
(Lev 2001). Kozberg (2009) showed the impact of current investment in R&D and
advertising in terms of the company’s future profitability in Fig. 4.1.
Noe and Parker (2005) demonstrate analytically that a company which wants to
participate in the World Wide Web should adopt an aggressive strategy by investing
massively in intangibles (e.g. R&D and advertising) with the aim of creating strong
barriers to entry for new competitors. This strategy—“winner takes all”—is char-
acterised by a very asymmetrical distribution, highly positively skewed, in accor-
dance with the Pareto distribution. In this context, this type of business tends to
register a high probability of bankruptcy, but surviving companies enjoy increasing
returns. Connolly and Hirschey (1984) also provide empirical evidence that the
statistical significance of the variables R&D and advertising is enhanced if the
company establishes a dominant position in the sector. This dominant position
enables the company to make continued investments in intangible assets, creating

high barriers to entry, and thus perpetuating its growth opportunities. Kotha (1998)
argues that the value of a website is greater when it engages a larger virtual
community, as this makes it a most attractive website. The higher the virtual
community, the greater the demand, and so the greater the potential for future
revenues, lower fixed costs and marginal units, which is reflected in a substantial
increase in the company’s future profitability. In this sense, the web traffic boosts
the Internet as an intangible asset valuable to the company.7
To measure the impact of intangible assets in the MVE of the net firms, Rajgopal
et al. (2003: 137) operationalize the concept of network based on Metcalfe’s law.
According to Metcalfe’s law, when (N) people are connected to a network, the
value of this network will be equal to N × (N − 1) = N2 − N. To measure this
variable, the authors use as a proxy the “number of unique visitors”. The network
variable is defined as
network ¼ ðnumber of visitorsÞ2 ðnumber of visitorsÞ ð4:6Þ
The authors conclude that there is a significant explanatory power of the network
variable and so include this variable in the OM (which they implicitly assume as a
proxy for the variable vt), showing a significant increase in the value of R2. They
also show empirical evidence that the network variable is positively and signifi-
cantly associated with future profits and sales (whose proxies were analysts’ esti-
mates), a result that in their opinion supports the existence of increasing returns to
scale associated with investments in R&D and advertising.
However, these results are not consensual. Using the methodology of Almond
and Koyck, Hand (2001a) concludes that most investments in R&D and advertising
are not value creators, as the net present value (NPV) associated with these projects
is negative. Value creation is identified only with the group of companies classified
in the higher percentiles with reference to the ratio of R&D over sales. Pioneers and
large firms have invested aggressively in intangibles, in clear agreement with the
model of Noe and Parker (2005),8 these results that indicate that the Internet sector
tends to follow the structure of highly concentrated market.
The noise that typifies the evaluation of this type of project/company derives
from the strong uncertainty associated with the estimation of future cash flows. In
most the cases, they were still very young companies with some technological
sophistication, who go to the capital market despite registering losses (the phe-
nomenon of positive valuation of losses) with a new idea, not yet tested and which
7
The numbers about the global growth of the internet population are impressive. Shiller (2000)
points to a value of 100 million users in 1999, rising to projections for 2003 to 177 million, an
estimated growth of 77%.
8
Rajgopal et al. (2003) recognized that the pioneers were the most successful. This evidence is
documented with the extension to the initial model introducing the interactive variable
“network*Amazon”, a company that Copeland et al. (2000) consider the symbol of the new
economy.
requires high up-front investment, so that the potential results are far from
guaranteed.
Alongside this uncertainty, and contrary to White et al. (1997) and Aboody and
Lev (2000), it is also worth to mention the strong information asymmetry between
insiders (managers) and outsiders (investors), with the consequent agency costs,
given the investment profile that characterizes this type of company’s investments
in the intangible assets. The information asymmetry mainly derives from three
factors: (i) the specificity of this type of investment, which hampers comparability
between companies; (ii) the absence of organized markets for transactions of this
type of asset (other than patents and licenses); and (iii) as these investments are
treated as costs in the year they are incurred, they are not subject to impairment tests
(i.e. comparison between cost of acquisition/production and market value), so no
later additional information is available to investors about their investments.
As a result of the noise in the estimation of future cash flows and information
asymmetry, there is great divergence in investors’ expectations about the value
created by this type of company. This divergence in expectations justifies the
seemingly paradoxical behaviour of the market, with some authors holding that the
technology-based companies tend to be undervalued by the market, while others
hold the opposite opinion, asserting the overestimation of such shares.
In the functional fixation hypothesis (Hall 1993; Hall and Hall 1993), the market
tends to undervalue this type of company because investors do not incorporate their
expectations. Given the short time horizon, all future benefits associated with
investments undertaken by these companies are mainly in intangibles. In the current
study, the noise in the evaluation of this type of company derives not only from the
strong uncertainty associated with the estimation of future cash flows (reflected in
the high volatility of shares of these companies), but also from the distortion
conveyed by the information in the financial statements on the profitability of these
companies, given the costs of treatment of the variables R&D and advertising.
To analyze the volatility of returns recorded by technology companies, Chan
et al. (2001) obtained a positive and statistically significant coefficient for the
variable R&D/sales. These authors find that firms, with a higher ratio R&D/sales
and poor performance in the previous three years (the period of analysis of port-
folios created), are also those that have registered the highest growth rates in liquid
results and provide the investors with abnormal returns, in line with the results
obtained by Lakonishok et al. (1994) and Fama and French (1992), designating
these companies by “glamour stocks” (i.e., companies with high ratios of
R&D/sales and MVE/BVE).
Ikenberry et al. (1995), Loughran and Ritter (1995) and Lakonishok and Lee
(2001) justify these abnormal returns based on the fact that the market tends first to
ignore these companies given their poor performance (treatment of investment in
intangibles as costs undervalues the short-term profitability), which for Chan et al.
(2001) reflects the existence of a clientele effect associated with this type of
company. Managerial support for successive investments in R&D and advertising is
a vote of confidence in the future profitability associated with these investments.
This vote of confidence is further highlighted by the great pressure to reduce
investment in these items, given its immediate impact on short-term profits. Lev and
Sougiannis (1996) and Chambers et al. (2002) also argue that the abnormal returns
offered by these companies are an additional premium, given the higher level of risk
that is inherent in them. Kothari et al. (2002) provide empirical evidence in this
sense. They conducted a research to explain the results obtained by
technology-based companies on the basis of investments in tangible assets, intro-
ducing as variables “debt”, “dimension”, “industry” and “growth”. They conclude
that the results expected from the investments in intangibles are much more
uncertain than expected results from investments in fixed assets. The estimated
coefficient for intangible assets exceeds the estimated coefficient of fixed assets by
about four times. According to the authors, this uncertainty accentuates the agency
costs between managers/shareholders (insiders) and creditors (outsiders), because in
case of bankruptcy, intangible assets, given their specificity and lack of organized
markets for their transaction, tend to lose all their value. However, the literature
reaches no consensus on the investor behaviour profile with reference to this group
of companies. For example, Jensen (1993) takes the opposite view, arguing that
investors tend to overestimate future abnormal results of technology-based com-
panies, particularly when reporting losses, citing the example of biotechnology
companies. See the specific case of high expectations created by the announcement
of a potential cure for a type of cancer.
Focusing the analysis on net firms, and following Jensen (1993), there was an
overestimation of the share price of these companies. The overstatement of the
prices of such securities results from: (i) the effect of public impression, i.e. the
media coverage, particularly in the 1990s, with a strong impact on the public, as
documented by Shiller (2000); (ii) the fact that the media tend to focus their
attention on successful pioneering companies (e.g. Yahoo and eBay); and (iii) the
intense marketing campaigns involving IPOs, particularly net firms (DuCharme
et al. 2001; Demers and Lewellen 2003; Jorion and Talmor 2006).
This suggests that the perception of high growth opportunities created by the
Internet led investors to extrapolate strong-growth expectations to all net firms,
based on the success of a small number of successful pioneers, for fear of “losing
the train of opportunity”. As demonstrated by Hugonnier et al. (2005) through a
general equilibrium investment model, the option of deferral given the type of
investment, such as investment in information technology, can erode the value of
this type of option, even in a moderate risk-aversion scenario. In short, for evalu-
ation, the literature is consensual, stressing that it is crucial to note the stage of the
business life cycle. Anthony and Ramesh (1992: 204) contend that:
At each stage of growth in an entity’s life cycle, different measures of financial performance
take on varying degrees of importance.
Thus, given the emerging nature of the Internet sector/net firms, we found that
the strategy of these companies followed a maximization of growth. Given the
profile of technology-based companies, investments are focused on the intangible
assets—R&D and advertising. Given the tax procedures set out by GAAP, losses
are a result of the effect of conservatism accounting. In this context, the unrecorded
goodwill (the difference between the MVE and the BVE) reflects the present value
(PV) of growth opportunities held by these companies and is enhanced by the
network effects created by the World Wide Web, not yet reflected in the financial
statements. Under the principle of rationality, it is to be expected that managers
only undertake investment projects that are associated with expectations of
abnormal returns.
The next chapter, in Part II, we initiate the results of our empirical study. Given
the magnitude and persistence of the losses reported by net firms, our goal is to
analyze the valuation of these companies by the market, or more specifically, to
examine the statistical relationship between market capitalization and reported
losses.
References
Abdel-Khalik AR (1975) Advertising effectiveness and accounting policy. Acc Rev 50:657–670
Aboody D, Lev B (2000) Information asymmetry, R&D, and insiders gains. J Finance 55(6):2747–
2766
Anthony JH, Ramesh K (1992) Association between accounting performance measures and stock
prices. J Acc Econ 15:203–227
Basu S (1997) The conservatism principle and the asymmetric timeliness of earnings. J Acc Econ
24:3–37
Ben-Zion U (1978) The investment aspect of nonproduction expenditures: an empirical test.
J Econ Bus 30:224–229
Chambers D, Jennings R, Thompson RB (2002) Excess returns to R&D intensive firms. Rev Acc
Stud 7:133–158
Chan SH, Martin JD, Kensinger JW (1990) Corporate research and development expenditures and
share value. J Financ Econ 26(2):225–267
development expenditures. J Finance 56(6):2431–2456
Chauvin KW, Hirschey M (1993) Advertising, R&D and the market value of the firm. Financ
Manage 22(4):128–140
Connolly RA, Hirschey M (1984) R&D, market structure, and profits: a value-based approach.
Rev Econ Stat 66(4):682–687
Damodaran A (2001) The dark side of valuation. Prentice Hall, United States
Econ 68(3):413–437
Diamond DW (1989) Reputation acquisition in debt markets. J Polit Econ 89:828–862
DuCharme LL, Rajgopal S, Sefcik SE (2001) Why was internet IPO underpricing so severe?
Working paper, University of Washington. http://ssrn.com/abstract=285986
Dukes R (1976) An investigation of the effects of expensing research and development costs on
security prices. Working paper. In: Proceedings of the conference on topical research in
accounting, New York University
Fama E, French K (1992) The cross-section of expected returns. J Finance 47:424–465
References 81
Garman M, Klass M (1995) On the estimation of securities price volatilities from historical data.
J Bus 81(5):67–78
Graham JR, Harvey CR (2001) The theory and practice of corporate finance: evidence from the
field. J Financ Econ 60(2–3):187–243
Hall BH (1993) The stock market’s valuation of R&D investment during the 1980s. Am Econ Rev
83(2):259–264
Hall BH, Hall RE (1993) The value and performance of US corporations. Brookings Papers Econ
Act 1:1–34
Hand JRM (2001a) Evidence on the winner-takes-all business model: the profitability
returns-to-scale of expenditures on intangibles made by U.S. internet firms, 1995–2001.
Working paper. Kenan-Flager Business School, Chapel Hill
Hand JRM (2001b) The role of book income, web traffic, and supply and demand in the pricing of
US internet stocks. Eur Finance Rev 5(3):295–317
(eds) Intangibles: an Oxford management reader. Oxford University Press, Oxford
Hirschey M (1982) Intangible capital aspects of advertising and R&D expenditures. J Ind Econ 34
(4):375–390
Ikenberry D, Lakonishok J, Vermaelen T (1995) Market underreaction to open market shares
repurchases. J Financ Econ 39:181–208
Jensen M (1986) Agency costs of free cash flow, corporate finance and takeovers. Am Econ Rev
76(2):383–393
Jensen M (1993) The modern industrial revolution, exit, and the failure of internal control systems.
J Finance 48:831–880
Jorion P, Talmor E (2006) Value relevance and non-financial information in emerging industries:
the changing role of web traffic data. Working paper, University of California, Irvine and
London Business School. http://ssrn.com/abstract=258869
Joos, P, Plesko, G. A. (2005) Valuing Loss Firms. Account Rev 80(3):847–870
Kotha S (1998) Competing on the internet: how amazon.com is rewriting the rules of competition.
Adv Strategic Manage 15:239–265
Kothari SP, Laguerre TE, Leone AJ (2002) Capitalization versus expensing: evidence on the
uncertainty of future earnings from capital expenditures versus R&D outlays. Rev Acc Stud 7
(4):355–382
accounting and non-financial information in explain revenues and valuations for internet firms.
VDM
Lakinishok J, Shleifer A, Vishny RW (1994) Contrarian investment, extrapolation, and risk.
J Finance 49:1541–1578
Lakonishok J, Lee I (2001) Are insider traders informative? Rev Finan Stud 14:79–111
Lev B (2001) Intangibles: management, and reporting. The Brookings Institution, Washington DC
Lev B, Sougiannis T (1996) The capitalization, amortization, and value-relevance of R&D.
J Account Econ 21(1):107–138
Lo K, Lys T (2001) The Ohlson model: contribution to valuation theory, limitations, and empirical
applications. J Acc, Auditing Finance 15(3):337–367
Loughran T, Ritter J (1995) The new issue puzzle. J Finance 50(1):23–51
Malkiel BG (1963) Equity yields, growth, and the structure of shares prices. Am Econ Rev 53
(5):1004–1031
Mao JCT (1966) The valuation of growth stocks: the investment opportunities approach. J Finance
21(1):95–102
firms. Paper presented at 14th annual conference on financial economics and accounting
(4):411–433
Moore M (2002) Intangible asset valuation using the Feltham–Ohlson framework and real option
analysis: theory and empirical evidence. Working paper. The State University of New Jersey
Hugonnier J, Morellec, E, Sundaresan SM (2005) Irreversible investment in general equilibrium.
Working paper presented at the 18th Australian finance and banking conference. University of
New South Wales, Sydney
Noe T, Parker G (2005) Winner take all: competition, strategy, and the structure of returns in the
internet economy. J Econ Manage Strategy 14(1):141
Ofek E, Richardson M (2002) The valuation and market rationality of internet stock prices. Oxford
Rev Econ Policy 18(3):265–287
Ofek E, Richardson M (2003) Dot.Com mania: the rise and fall of internet stocks prices. J Finance
58(3):1113–1138
Ohlson J (1995) Earnings, book values, and dividends in equity valuation. Contemp Acc ev 11
(2):661–687
Peles Y (1971) Rates amortization of advertising expenditures. J Polit Econ 79:1032–1058
Picconi M (1977) A reconsideration of the advertising assets on financial statement. J Acc es
15:317–326
Quintar A, Zisswiller R (1994) Financial theory. Biblioteca de Economia e Gestão, Caminho
Rajgopal S, Venkatachalam M, Kotha S (2003) The value relevance of network advantages: the
case of e-commerce firms. J Acc Res 41(1):135–162
Finance 44(2):223–333
Sougiannis T (1994) The accounting based valuation of corporate R&D. Acc Rev 69(1):44–68
Szewczyk SH, Tsetsekos GP, Zantout Z (1996) The valuation of corporate R&D expenditures:
evidence from investment opportunities and free cash flow. Financ Manage 25(1):105–110
Taylor W (1974) A note on Mao´s growth stock-investment opportunities approach. J Finance 29
(5):1573–1576
Weiss LW (1969) Advertising, profits, and corporate taxes. Rev Econ Stat 51:421–430
Wiley, New York
Wooldridge R (1988) Competitive decline and corporate restructuring: is a myopic stock market to
blame? J Appl Corp Finance 1:26–36
Zantout Z, Tsetsekos G (1994) The wealth effects of announcements of R&D expenditures
increases. J Financial Res 17(2):205–216
Part III
Empirical Study
Chapter 5
Period, Sample Selection and Definition
of Variables
Abstract This chapter presents the definitions of the basic concepts underlying the
empirical study. We begin by defining the period investigated, which we define as
the New Economic Period (NEP) as well as the concept of net firms. Then, we
define the procedures to select the sample. We present the number of observations
and also the systematization of the investigation. For comparative purposes, we
define also a match sample—no-net firms and finally we present the definition of
the variables used.
Keywords Empirical study New economy period (NEP) Net firms Control
sample—non-net firms
5.1 Introduction
This chapter introduces the core concepts underlying the empirical study.
Therefore, and due to the emerging nature of the Internet sector, we begin by
specifying the period that will be investigated, which we call the New Economy
Period (NEP) and the profile of the companies in the new economy—net firms. We
also explain the procedures to select a match sample, i.e. non-net firms with an IPO
contemporaneous to the net firms. With the selection of this match sample, we aim:
(i) to control the effect of fashion (fad) associated with net firms documented by Lee
(2001), Cooper et al. (2001) and further (ii) to compare results between the two
samples in order to ensure robustness of the results and conclusions, since,
according to Bartov et al. (2002), the phenomenon of “positive valuation of losses”
seems to have assumed greater importance in the universe of net firms.
After selecting the two samples, we made a comparative analysis, particularly in
light of the number of IPOs per year, and the partition of the companies on the basis
of results achieved—profit or loss, movements on the market. Surprisingly, we
found that the failure rate in both samples was very low, as opposed to the rate of
mergers and acquisitions (M&A), which exceeded 80%. Further emphasis was done

DOI 10.1007/978-981-10-3009-3_5
86 5 Period, Sample Selection and Definition of Variables
to the comparison in terms of the industry. Finally we define the variables used in
the study. The database used is the Compustat North America.
5.2 Definition of the “New Economy Period—NEP”
One of the most important aspects of this research is to identify the Period of the
New Economy (NEP) and hence define the object of study—new economy com-
panies, which for brevity we will denote by the term net firms.1
It is difficult to identify the beginning of the NEP. Core et al. (2003: 54) state that
the concept NEP first appears in Dow Jones News Service when it mentions that in
1994 the bank attributed to Nova Scotia’s “old economy” a growth rate of 3%. On
09 May 2000, the Chairman of the U.S. Federal Reserve commented that economic
growth had accelerated in the period of the new economy. In 1995 appeared the first
browser from Netscape Communications, and Core et al. (2003) referred to
Amazon the symbol of the new economy.
Thus, we identified the early period of the new economy (NEP) with the year
1996. The period of this research covers the years 1996–2003.
5.3 Definition of “Net Firms”
The definition of new economy company—net firm—also proves to be highly

complex. The sector of activity (Standard Digital Code—SIC) is of little use, since
this type of company tends to operate in different sectors, even considering the
substantial arbitrariness of criteria.
However, the difficulty of defining a net firm is overcome by the creation of
specific market indexes for these companies. Thus, for this research we consider net
firms, companies listed in the Internet Stock List index (ISDEX).2 However, the
identification of net firms based only on this index brings us the problem that in our
sample we include only surviving firms (survivor bias effect). Therefore, the
empirical study should control for this effect.
1
The designation itself is not consensual. For example, Neves (2002: XVI) prefers the designation
“Digital Economy”. In the preface to this same book, Brealey opts for the concept “New
Economy”. In the present investigation we chose the concept “New Economy”, which we des-
ignate by the acronym NEP as it was the concept with greater generalization (Copeland et al. 2000;
Damodaran 2001; Penman 2001; Jorion and Talmor 2006; Kaplan 2002; Core et al. 2003).
2
This index was originally created by the company WRSN—Wall Street Research Net, posted on
the website http://www.wsrn.com, for the period 20 April 1999 to 11 November 2003. At the end
of 2003, WSRN was acquired by the Internet.com company. Currently on the website http://www.
bullsector.com/internet.html can be found multiple indexes on net firms [GSTI—Internet Index
(WBC: ^ GIN), AMEX—Internet Components Index (AMEX: ^ IIX), Philadelphia Internet Index
(STREET.COM: ^ DOT), U.S. Dow Jones (DJI: ^ DJUSNS)].
5.4 Criteria for Selection of Samples 87
5.4 Criteria for Selection of Samples
5.4.1 New Economy Companies: “Net Firms”
In order to overcome the problem of the survival effect, we began by selecting a

sample of net firms listed in the ISDEX index on 27 June 2003. Based on this date
we obtained a list of 242 companies. Next, we analysed the reports of Morgan
Stanley and (i) “The Technology IPO Yearbook—8th Edition” with a list of 286 net
firms quoted on 6 March 2002, (ii) “The Technology IPO Yearbook—7th edition”
with the 382 listed companies on 21 February 2001, (iii) “The B2B Internet Report”
also of Morgan & Stanley, with 354 listed on 24 April 2000 and (iv) companies
listed in “The Internet Company Handbook” also of Morgan & Stanley, with 108
listed on 25 May 1999. Based on these reports, it was possible to collect infor-
mation about net firms listed before and after the crash that occurred in the first
quarter of 2000.3 Comparing our sample with Jay Ritter´s sample of net firms, we
identified 534 companies on 2002.4 Finally, we analysed the various investigations
into these companies (see Table 5.2).
Tables 5.1 and 5.2 summarize this.
With this procedure, we initially identified 658 companies. A total of 24 com-
panies have been excluded from this group because of a lack of information, i.e. the
companies did not have their information tabulated in the Compustat database on
the three key variables for analysis: “market capitalization—MVE”, “value of
equity—BVE” and “net income”. We have also further eliminated 12 companies
that were identified as outliers.
The identification of outliers deserves particular attention, because, based on the
analysis of boxplots with reference to the variables “market capitalization—MVE”,
“equity—BVE” and “net income”, the outliers identified were numerous. In order
to preserve the representativeness of the sample and ensure the robustness of the
results, we chose to identify outliers according to the analysis of “inner fences” and
“outer fences” indicators calculated by Eviews (Eviews Manual 2004: 397).5
Graphically this is expressed in Fig. 5.1.
After this purification process of the data, the final sample of net firms comprises
622 companies. This sample is one of the most representative of the new economy
firms, compared with other studies (see Table 5.2), both in size, where the survival
effect is taken into account, either with reference to the period under study, 8 years,
and covering both periods: the pre-crash period (1996—first quarter of 2000) and
3
The crash occurred in the firstt quarter 2000; Demers and Lev (2001) documented a decrease of
45% in the ISDEX index for the period February to March 2000.
4
This database can be found on the homepage of Jay Ritter: http://bear.cba.ufl.edu/ritter/ipodata.
htm.
5
Analytically: (i) inner fences = [1stQ – 1.5 * IQR; 3rdQ * +1.5IQR] and (ii) outer
fences = [1stQ – 3rdIQR; 3rdQ + 3IQR], with Q-quartile and IQR—the difference between
quartile (inter quartile range).
Table 5.1 New economy firms listed on ISDEX

ISDEX 27/06/2003 06/03/2002 21/02/2001 24/04/2000 25/05/1999 Jay
Ritter
Number of 242 286 382 354 108 534
companies
the post-crash period (second quarter 2000–2003) of the “dot.com bubble” (Demers
and Lev 2001).
Indeed, one of the consensual aspects in the various studies on the subject of
evaluation of net firms lies in the need to consider longer periods of analysis, given:
(i) the emerging nature of the Internet industry (life cycle perspective), and espe-
cially (ii) the results for the period 1999–2000, which researches cover (see
Table 5.2), should be analysed with “great caution” given the effect of the “dot.com
bubble” (Penman 2001; Talmor 2001; Lewellen 2003).6
5.4.2 Companies with a IPO Date Contemporaneous

to “Net Firms”: “Non-Net Firms”
The 1990s were clearly a period of “hot markets”, a phenomenon widely docu-
mented in the literature on IPOs.7 Loughran and Ritter (2003) documented a total of
6169 IPOs for the period of 1980–2000, 14% concentrated in the period 1990–1998
and 65% in 1999–2000. Ljungvist and Wilhelm (2003) present similar results for
the period 1996–2000.
Thus, taking into account the cluster effect that the IPOs tend to register and
particularly the results obtained by Bartov et al. (2002), who documented the effect
of “positive valuation of the losses” only for net firms, we also select a sample of
contemporaneous IPOs of non-net firms for comparison purposes (match sample),
which we will designate by non-net firms. The use of a control sample has been
revealed to be very useful in order to corroborate the results obtained for the main
sample, despite limitations in their selection.
To select this sample, we used the database provided by the National
Association of Securities Dealers Automated Quotation (NASDAQ), which con-
tains all IPOs that occurred between 1990 and 2002 (until September) in the
NASDAQ index.8 Thus, we began by identifying in this database the net firms we
had already selected. The next step was to identify non-net firms contemporary with
6
The dot.com bubble, characterized by prices in a given period of time that diverge from funda-
mental values, was documented for example by Ofek and Richardson (2002, 2003), Ljungvist and
Wilhelm (2003), Loughran and Ritter (2003), and Keating et al. (2003).
7
As an example we quote: Ritter (1991) and Loughran and Ritter (1995).
8
Thanks to Darren Hawkins, the department NASDAQ International for helping with this database.
Table 5.2 Sample sizes of net firms: systematization of other researches

Authors Number of Period of Source of information
companies analysis
Core et al. (2003) 214 1999 Sample Hand (2001a) (Ψ)
Rajgopal et al. 92 1999Q1–2000Q3 Internet stock list
(2003)
Keating et al. 148 Two dates: Internet stock list
(2003) • 13/03/00
• 05/26/00
Loughran and 534 1990–2000 • SEC—Securities and Exchange
Ritter (2003) Commission
• CRSP—Centre of Research Security
Prices
Ljungqvist and 552 01/1990– Sample Jay Ritter, added 18 companies
Wilhelm (2003) 02/2000
Demers and 373: Post 02/1999 Internet stock list
Lewellen (2003) • 191 B2C
• 182 B2B
Hand (2003) 274 1997Q1–2000Q3 Internet stock list
Ofek and 305 1998Q1–2000Q4 Reports and Morgan Stanley
Richardson
(2003)
Bartov et al. 98 1996Q1–1999Q3 Internet stock list
(2002)
Martinez and 245 1996Q1–2001Q2 Internet stock list
Clement (2002)
Bowen et al. 174 1999 Report annual accounts (10-k)
(2002)
Rajgopal et al. 45 (Only Up to 09/2002 The B2B internet report and Morgan
(2002) B2B) Stanley
Davis (2002) 273 Up to 08/2000 Internet stock list
Tokic (2004, 518 initial 1996–2000 All companies belonging to the two
2005) sample digits GIS industry sub-sector
Ofek and 400 1999 Reports and Morgan Stanley
Richardson
(2002)
Liu and Song 369 02/2000 Internet stock list
(2001)
Hand (2001a) 274 1997Q1–1999Q2 Internet stock list
Hand (2001b) 285 1995Q1–2001Q2 Internet stock list
Bagnolli et al. 98 31/03/99– First call forecast analysts
(2001) 31/10/00
Demers and Lev 84 (B2C Three dates: Internet stock list
(2001) only) • 3/12/1999
• 28/02/2000
• 31/05/2000
Trueman et al. 95 1998Q4–2000Q2 Internet stock list
(2001)
(continued)

Authors Number of Period of Source of information
companies analysis
Kozberg (2009) 316 02/1999– Internet stock list
05/2001
DuCharme et al. 342 1990–1999 Internet stock list
(2001)
Hendershott 435 1995–2000 Venture Xpert Database (Thompson
(2001) financial)
Schultz and 420 1996 to March Internet stock list
Zaman (2001) 2000
Trueman et al. 63 09/1998 and Internet stock list
(2000) 12/1999
Q—Information collected on a quarterly basis
B2B—Business–to–Business
B2C—Business–to–Consumer
Ψ—Sample the first version of a working paper “Profit, Losses and Non-Linear Pricing of Internet
Stocks”
Ø Far Outliers
OUTER FENCES
Close Outliers
INNER FENCES
3rd Quartile
Mean
* Median
1st Quartile
Eviews (2004:version 5.0: 397).

Fig. 5.1 Identification of outliers
the dates of the IPO of the net firms. The initial day differential was 3–5 days. For
the period before 1999, the maximum differential was 30 days like Schultz and
Zaman (2001), given that a lower number of IPOs occurred during this period. For
the years 1999 and 2000, it was even more difficult to find “a match firm” given the
predominance of net firm IPOs. This also made it impossible to control the industry
effect, i.e. the selection of non-net firms contemporaneous with the date of the net
firm IPOs and belonging to the same sector (SIC).
The initial sample includes 564 companies. As the sample of net firms, the sample
was subjected to a purification of data process, and 11 companies have been elimi-
nated due to missing information. Twelve companies were classified as outliers
according to the same criteria for net firms, thus having been removed for the entire
period of analysis, i.e. 1996–2003. The final sample comprises 541 non-net firms.
In the following section, we analyse the composition of the two samples.
Comparisons have also been established based on the results recorded by the
companies in each sample, movements of entries and exits from the market, and
sectors and markets where they are predominantly listed.
5.5 Composition and Comparative Analysis of the Two

Samples: “Net Firms” and “Non-Net Firms”
5.5.1 The Samples of “Net Firms” and “Non-Net Firms”
Comparing the size of the two samples net firms (Table 5.3) and non-net firms
(Table 5.4), with the total number of IPOs that occurred in three U.S. markets:
NASDAQ, New York Stock Exchange (NYSE) and American Stock Exchange
(AMEX) (Table 5.5), the two samples selected are representative of the number of
IPOs that took place in the three markets. For example, in 1999 the number of IPOs
accounted for 432 (279 net firms and 153 non-net firms), when the number of IPOs
that took place on NASDAQ was 485 (see Table 5.5).
5.5.2 Comparative Analysis of “Net Firms” Versus “Non-

Net Firms”
Comparing the two samples (Tables 5.3 and 5.4), we confirm a greater number of
IPOs before the year 1996 for the group of non-net firms (26.43%), against 10.93%
reported by net firms. Both samples highlight the “cluster” effect in time, verifying a
strong concentration of the IPOs during the “dot.com bubble”, i.e. the years of 1999
and 2000. The group of net firms, however, experienced a higher percentage—
69.30% (44.86 and 24.44%) against 50.28% (28.28 and 22.00%) for the group of
non-net firms. This preliminary result supports the “usefulness” of resorting to a
control sample for comparative purposes. In the period after 2000, there are few
IPOs in both groups, in line with the overall behaviour of the three markets:
NASDAQ, NYSE and AMEX (Table 5.5).
With reference to the percentage of firms that left the market, the results are
similar: 33.76% (210/622) in the group of net firms compared with 27.54%
92
Table 5.3 Composition of the sample of net firms

Year of Global Split the sample No. IPO % Results reported by the Reason left the market
IPO sample by year IPO companies on the date left the
marketb
Profits Losses Profits Losses Total M&Ac Bankruptcy Liquidation Private Othersd
≤1996 71 34 (48%) 37 (52%) 68 10.93 0 0 0
1997 115 49 (43%) 66 (57%) 49 7.88 0 0 0
1998 154 47 (31%) 107 (69%) 56 9.00 2 2 4 4
1999 430 80 (19%) 350 (81%) 279 44.86 12 1 13 12 1
2000 527 74 (14%) 453 (86%) 152 24.44 37 4 41 39 2
2001 428 36 (8%) 392 (92%) 7 1.13 59 9 68 60 1 2 5
2002 367 54 (15%) 313 (85%) 3 0.48 34 2 36 30 3 3
2003 307 101 206 (67%) 1 0.16 43 5 48 34 1 4 2 7
(33%)
s/data 7a 1.13
IPO
Total 2399 622 187 23 210 179 3 (1.43%) 9 (4.29%) 2 17
(89.05%) (10.95%) (85.24%) (0.95%) (8.10%)
The composition of the sample by year does not correspond to the formula: global sample + number of IPOs—number of firms that left the market, because some firms
left the market in year t + 1 but we only have data to the year t. So, we consider that the firm left the market in the year t
a
It was not possible to identify the date of the IPO
b
If the company report data for the year t but left the marker on year t + 1, we consider that the firm left the market on date t
c
M&A—Mergers and Acquisitions
d
For example failed to obtain information from Security Exchange Commission (SEC)
5 Period, Sample Selection and Definition of Variables
Table 5.4 Composition of the sample of non-net firms (see footnote of Table 5.3)
Year of Global Split the sample No. IPO % Results reported by the Reason left the market
IPO sample by year IPO companies at the date left the
marketb
Profits Losses Profits Losses Total M&Ac Bankruptcy Liquidation Private Othersd
≤1996 143 81 (57%) 62 (43%) 143 26.43 2 0 2 2
1997 191 109 (57%) 82 (43%) 60 11.09 1 2 3 3
1998 212 120 (57%) 92 (43%) 52 9.61 7 2 9 9
1999 356 182 (51%) 174 (49%) 153 28.28 14 11 25 20 1 4
2000 445 195 (44%) 250 (56%) 119 22.00 19 13 32 24 4 4
2001 402 140 (35%) 262 (65%) 5 0.92 16 9 25 22 1 2
2002 374 145 (39%) 229 (61%) 4 0.74 16 7 23 17 2 4
2003 336 146 (43%) 190 (57%) 0 0.0 20 10 30 25 5
No. IPO 5a 0.92
date
Total 2459 541 95 54 149 122 5 (3.36%) 0 (0.00%) 3 19
(63.76%) (36.24%) (81.88%) (2.01%) (12.75%)
5.5 Composition and Comparative Analysis of the Two Samples …
a
It was not possible to identify the date of the IPO
b
If the company report data for the year t but left the market on year t + 1, we consider that the firm left the market on date t
c
M&A—Mergers and Acquisitions
d
For example failed to obtain information from Security Exchange Commission (SEC)
93
Table 5.5 The number of IPOs by market and by year

Year NASDAQ NYSE AMEX Total
Total %c Total %c Total %c Total %
1992 442 10.26 80 10.94 6 4.88 528 10.23
1993 520 12.07 97 13.27 11 8.94 628 12.16
1994 444 10.30 82 11.22 13 10.57 539 10.44
1995 476 11.05 72 9.85 9 7.32 557 10.79
1996a 680 15.78 88 12.04 18 14.63 786 15.22
1997 494 11.46 87 11.90 22 17.89 603 11.68
1998 273 6.34 68 9.30 21 17.07 362 7.01
1999 485 11.26 49 6.70 11 8.94 545 10.56
2000 397 9.21 48 6.57 6 4.88 451 8.74
2001 63 1.46 35 4.79 3 2.44 101 1.96
2002b 35 0.81 25 3.42 3 2.44 63 1.22
Total 4309 83.46c 731 14.16c 123 2.38c 5163c
a
The begin of the New Economy Period (NEP)
b
Information until September 2002
c
Percentage of total IPOs (i.e. NASDAQ, NYSE e AMEX: total 5163) (http://www.nasdaq.com;
http://www.nyse.com and http://www.amex.com)
(149/541) in the group of non-net firms. As for the exit reason, the results are
interesting. In both samples we find that the main reason for leaving the market is
due to mergers and acquisitions (M&A), in this case, more than 80%. The failure
rate, in turn, is less than 5%, results that contradict our initial expectations.
Given the scale of the crash, with a fall of over 40% in the ISDEX index during
the first quarter of 2000 (Demers and Lev 2001; Ofek and Richardson 2002, 2003;
Keating et al. 2003), we expected to obtain a higher rate of bankruptcies, at least in
the group of net firms. Surprisingly this group showed a lower failure rate (1.43 vs.
3.36% in the group of non-net firms).
The high percentage of M&As is justified, because the Internet is an emerging
sector. Thus, initially attracted by the high share capitalizations, new firms enter in
the market.9 With the increased competition, and according to Kaplan (2002), we
are witnessing a process of rationalization of this sector.
This empirical evidence seems to confirm the “winner-takes-all” strategy which
Noe and Parker (2005) showed mathematically and which Hand (2001a),
Hendershott (2001) demonstrated empirically. According to the model of Noe and
9
Hendershott (2001) uses the term “imitation” to describe the IPO boom of net firms that occurred
in 1999 and 2000. But the interesting fact according to the data of this author, and in line with the
results reported by Schultz and Zaman (2001), is that this group of companies (imitators) could
attract funding from venture capital as well as the involvement of the most prestigious investment
banks (with a reputation to uphold) to the IPO process. Schultz and Zaman (2001) reported that the
six major investment banks (CS First Boston, Deutsche Bank, Goldman Sachs, Merrill Lynch,
Morgan Stanley and Salomon Brothers) were associated with 39.30% of the net firms IPO (un-
derwriter) in the period 1996 to March 2001, against 27.40% of other contemporaneous IPOs.
5.5 Composition and Comparative Analysis of the Two Samples … 95
Table 5.6 Composition of net firms and non-net firms by industry

SIC Standard Industrial Code (SIC) Net firms Non-net firms
Total %d Total %d
2834 Pharmaceutical preparationsa 1 0.16 22 4.07
2836 Biological products (no diagnostic substances)a 17 3.14
3576 Computer communication equipmenta 17 2.73 6 1.11
3577 Computer peripheral equipment, NECa 2 0.32 8 1.48
3661 Telephone and telegraph apparatusa 13 2.09 10 1.85
3663 Radio & TV broadcasting and communications 10 1.61 8 1.48
equipmenta
3674 Semiconductors and related devicesc 7 1.13 33 6.10
3826 Laboratory analytical instrumentsc 8 1.48
4812 Radiotelephone communicationsa 3 0.48 9 1.66
4813 Telephone communications (no 24 3.86 18 3.33
radiotelephone)c
4832 Radio broadcasting stationsa,b 2 0.32 9 1.66
4841 Cable and other pay television servicesb 9 1.66
4899 Communications services, NECa,b 7 1.13 6 1.11
5961 Retail—catalog and mail-order housesa 31 4.98
7370 Services—computer programming, data 183 29.42
processing, etc.c
7371 Services—computer programming servicesc 8 1.29
7372 Services—pre-packaged softwarec 160 25.72 68 12.57
7373 Services—computer integrated systems 34 5.47 14 2.59
designa,b
7374 Services—computer processing and data 2 0.32
preparationc
7389 Services—business services, NECa 13 2.09
8731 Services—commercial physical and biological 13 2.40
researchc
Total 517 83.12 258 47.69
Industrial sector classified as high technology according the criteria of:
a
Francis and Shipper (1999)
b
Collins et al. (1997)
c
Loughran and Ritter (2003)
d
Percentage calculated based on the total companies by sample (622 net firms; 541 non-net firms)
Parker (2005), the pioneers that invest massively in intangibles, such as R&D and
advertising, were the companies that survived. The strategy is to achieve high
market shares, with the purpose of generating the network effect made possible by
the Internet that enabling them to achieve increasing returns.10
10
With this reasoning, Noe and Parker (2005) also support, albeit indirectly, the phenomenon of
“positive valuation of losses” based on the “conservatism accounting” modelled by the FOM. At
Table 5.7 Number of companies listed by market

Market Net firms Non-net firms Total
Total % Total % Total %
NYSE—New York Stock Exchange 11 1.77 1 0.18 12 1.03
AMEX—American Stock Exchange 3 0.48 2 0.37 5 0.43
NASDAQ 505 81.19 482 89.09 987 84.87
Regional stock exchange 1 0.16 1 0.09
LBO—Leverage Buy Out 18 2.89 8 1.48 26 2.24
Ohers 84 13.51 48 8.88 132 11.35
Total 622 100 541 100 1163 100
Table 5.6 show that out of a total of 158 sectors, approximately 47.69% of non-
net firms companies are classified as high-tech companies, according to the criteria
defined by Collins et al. (1997), Francis and Schipper (1999), Loughran and Ritter
(2003), sectors to some extent which are also emerging. This effect is not differ-
entiated between the two samples. This emerging character becomes stronger in net
firms, since the concentration of firms in high-technology sectors is even greater.
Dispersing the aggregate sample by 74 sectors, the percentage of high-tech com-
panies amounted in this group to 83.12%.11
Unsurprisingly, and in accordance with the overall market behaviour, we can see
in Table 5.7 that more than 80% of net and non-net firms are listed on NASDAQ.
In the next section, we define the variables used in the empirical study, selected
from the Compustat North America database.
5.6 Definition of Variables
The database used was Compustat—version Compustat North America. To collect

data we chose to use information on an annual basis because: (i) this information
corresponds to information already available after all corrections made to the
financial statements (restated financial statements), and (ii) for the variable “ad-
vertising” (which corresponds to the description Advertising), the Compustat
database only provides information on an annual basis. This variable is a key
variable in this research, as it acts as proxy for growth opportunities for the group
net firms, whose business model is B2C—“Business to Consumers” (see criteria for
subdivision of the sample in Sect. 6.3). Choosing to work with quarterly data would
involve manual collection of the data for this variable for both samples under study.
the start-up/growth phase and in order to create future growth opportunities, those firms invest
massively in intangible assets which were accounted as costs under GAAP.
11
See Appendixes 5.1 and 5.2 with the number of net firms and non-net firms by sector.
5.6 Definition of Variables 97
Also note that given that the terminus of a fiscal year may occur at different
months, the data collection was made based on the “calendar year basis”.12
Activating this option, the database provides information that it is comparable
between companies and in different periods of the year. This option also corrects the
information of the effect of stock splits and dividends. As the unit of measurement,
quantitative variables have been measured in thousands of dollars. Thus, in order to
undertake the empirical study, information was collected for the variables as listed
in Table 5.8.
We defined the variable income before extraordinary items—available for
common (Res_IExt), and according to the manual of the Compustat data base
(2004: 249), as corresponding to adjusted net income from extraordinary items
(item Annual 192), discontinuation of operations, e.g. the result obtained with the
closure of a given division (item Annual 66), less preferred dividends and the effect
of change in accounting policies (Compustat Manual 2004).13
With reference to the R&D and advertising variables, and like other studies (e.g.
Fama and French 1998; Core et al. 2003) outside the universe of net firms and Hand
(2001b, 2003) in the universe of net firms, when the information is not available
(NA—not available), these variables assume a null value. Although with this
procedure, we are underestimating the values of these variables, this option allows
us to “preserve” the sample size. Contrary to usual practice, we include in the
sample firms that have negative values for the variable BVE.14
Since the universe of companies under analysis is characterised by firms in the
start-up/growth stage, negative values for the variable BVE reflect the effect of
conservatism accounting (Zhang 2000). Consecutive investments in R&D and
advertising, tend to understate net income, with cumulative effects on the variables
assets and BVE. Thus, a negative value for the variable BVE reflects the need for
additional investments in operating assets (Zhang 2000). On the other hand,
Damodaran (2001) argues that until the date of maturity of the debt, the purchase
option (call option) owned by shareholders continues to have value, as far as the
debt firm value may increase. Thus, when there is high volatility in firms15 the risk
can be an ally of the shareholder. By undertaking new investment projects,
investors increase the likelihood of obtaining high yields if the investment is
12
With reference to the two samples analysed, net firms and non-net firms, 53.88% (54.34%) of
companies wound up the fiscal year in December, in March 4.66% (2.96%), in June 3.38%
(4.99%) and in September 41% (4.34%), respectively. Note that for 35.67% of net firms and
33.37% of non-net firms, information about the closing month of the fiscal year was not available
in the Compustat database.
13
For a more detailed analysis of the preparation of Financial Statements according to GAAP, see
Penman (2003).
14
Collins et al. (1997, 1999), Burgstahler and Dichev (1997), Barth et al. (1998, 2003), Tan
(2004), Joos and Plesko (2005) outside the universe of net firms and Hand (2001b, 2003) and
Trueman et al. (2000), in the universe of net firms, excluded from the sample under analysis
companies with a negative BVE.
15
The high volatility patterns recorded for net firms are documented by Ofek and Richardson
(2002, 2003).
Table 5.8 Variables Definition

Variables Acronyms Annual item
Balance sheet
Property, plant and equipment—total net value AF 8
Current assets—total AC 4
Receivables—total DR 2
Cash and equivalents CTN 1
Current assets—others OAC 68
Assets—total AT 6
Common equity—total BVE 60
Retained earnings LR 36
Long-term debt—total PMLP 9
Current liabilities—total PCP 5
Liabilities—total PT 181
Profit and loss account
Depreciation and amortization AMORT 14
Selling, general, and administrative expenses CGA 189
Cost of goods sold CMVMC 41
Advertising expenses PUB 45
Research and development expenses I&D 46
Special items IExt 17
Income before extraordinary items—available for common Res_IExt 237
Net income (loss)/profit RL 172
Sales (net) VENDAS 12
Statement of cash flows
Operating activities-net cash flows CFO 308
Others
Market value of equity MVE MKVALF
Cash dividends—common DIV 21
successful, and restrict their losses if the project fails (Myers 1977). Thus, elimi-
nating these firms would imply in our view a distortion of the results to be obtained.
Core et al. (2003: 51) conclude that by eliminating firms with negative equity in
their sample, they removed about 7.5% of the companies in the start-up phase and
5.6% of high-tech enterprises.
In summary, we conclude that the differences are not significant between the two
samples: indeed (i) both samples show a cluster effect in time of the number of
IPOs that occurred, particularly in the NASDAQ market; (ii) the percentage of firms
that exited the market was similar; (iii) the M&A is the main factor behind market
exit; and (iv) both samples tend to operate in high-tech sectors. The findings also
show that the companies that exit the market are those that reported the worst
performance, i.e. the percentage of firms reporting losses. In the next chapter, we
present the method employed in the study.
Appendix 5.1 Number of Net Firms by SIC 99
Appendix 5.1: Number of Net Firms by SIC
SIC Description No. firms

1040 Gold and silver ores 1
2721 Periodicals: publishing and printing 1
2741 Miscellaneous publishing 1
2750 Commercial printing 5
2834 Pharmaceutical preparations 1
2844 Perfumes, cosmetics and other toilet preparations 1
3089 Plastics products, NEC 1
3541 Machine tools, metal cutting types 1
3571 Electronic equipment 1
3576 Computer communication equipment 17
3577 Computer peripheral equipment, NEC 2
3578 Calculating and accounting machines (no electronic computers) 1
3651 Household audio and video equipment 1
3661 Telephone and telegraph apparatus 13
3663 Radio and TV broadcasting and communications equipment 10
3669 Communication equipment, NEC 2
3670 Electronic components and accessories 1
3674 Semiconductors and related devices 7
3690 Miscellaneous electrical machinery, equipment and supplies 1
3714 Motor vehicle parts and accessories 1
3825 Instruments for measuring and testing of electricity and electrical signals 1
3990 Miscellaneous manufacturing industries 1
4700 Transportation services 6
4812 Radiotelephone communications 3
4813 Telephone communications (no radiotelephone) 24
4822 Telegraph and other message communications 3
4832 Radio broadcasting stations 2
4899 Communications services, NEC 7
4955 Hazardous waste management 1
5045 Wholesale—computers and peripheral equipment and software 3
5065 Wholesale—electronic parts and equipment, NEC 1
5122 Wholesale—drugs, proprietaries and druggists’ sundries 2
5400 Retail—food stores 1
5735 Retail—record and prerecorded tape stores 1
5812 Retail—eating places 1
5912 Retail—drug stores and proprietary stores 1
5940 Retail—miscellaneous shopping goods stores 1
5961 Retail—catalog and mail-order houses 31
5990 Retail—retail stores, NEC 3
(continued)
(continued)
SIC Description No. firms
6035 Savings institutions, federally chartered 1
6036 Savings institutions, not federally chartered 1
6162 Mortgage bankers and loan correspondents 2
6163 Loan brokers 1
6211 Security brokers, dealers and flotation companies 7
6282 Investment advice 1
6411 Insurance agents, brokers and services 4
6531 Real estate agents and managers (for others) 1
6794 Patent owners and lessors 1
7310 Services—advertising 7
7311 Services—advertising agencies 1
7320 Services—consumer credit reporting, collection agencies 1
7330 Services—mailing, reproduction, commercial art and photography 2
7331 Services—direct mail advertising services 5
7361 Services—employment agencies 1
7370 Services—computer programming, data processing, etc. 183
7371 Services—computer programming services 8
7372 Services—pre-packaged software 160
7373 Services—computer integrated systems design 34
7374 Services—computer processing and data preparation 2
7377 Services—computer rental and leasing 2
7380 Services—miscellaneous business services 1
7385 Services—telephone interconnect systems 1
7389 Services—business services, NEC 13
7812 Services—motion picture and video tape production 1
7841 Services—video tape rental 1
7990 Services—miscellaneous amusement and recreation 2
8200 Services—educational services 1
8600 Services—membership organizations 1
8700 Services—engineering, accounting, research, management 1
8711 Services—engineering services 1
8741 Services—management services 2
8742 Services—management consulting 9
9995 Non-operating establishments 1
Total 622
Fonte: http://www.sec.gov/info/edgar/siccodes.htm (U.S. Securities Exchange Commission)
Appendix 5.2 Number of Non-Net Firms by SIC 101
Appendix 5.2: Number of Non-Net Firms by SIC
SIC Description
100 Agricultural production-crops 1
1220 Bituminous coal and lignite mining 1
1311 Crude petroleum and natural gas 2
1382 Oil and gas field exploration services 1
1520 General bldg contractors—residential 1
1600 Heavy construction other than bldg const—contractors 1
1700 Construction—special trade contractors 1
2020 Dairy products 1
2040 Grain mill products 2
2086 Bottled and canned soft drinks and carbonated waters 1
2090 Miscellaneous food preparation and kindred products 1
2253 Knit Outerwear mills 1
2330 Women’s, misses’, and juniors outwear 1
2340 Women’s, misses’, children’s and infants’ undergarments 1
2390 Miscellaneous fabricated textile products 1
2522 Office furniture (no wood) 1
2650 Paperboard containers and boxes 1
2741 Miscellaneous publishing 2
2810 Industrial inorganic chemicals 1
2833 Medical chemicals and botanical products 2
2834 Pharmaceutical preparations 22
2835 In vitro and in vivo diagnostic substances 8
2836 Biological products (no diagnostic substances) 17
2844 Perfumes, cosmetics and other toilet preparations 1
2870 Agricultural chemicals 1
2890 Miscellaneous chemical products 1
3089 Plastics products, NEC 1
3140 Footwear (no rubber) 1
3270 Concrete, gypsum and plaster products 1
3290 Abrasive, asbestos and miscellaneous non-metallic mineral products 1
3312 Steel works, blast furnaces and rolling mills (coke ovens) 1
3350 Rolling drawing and extruding of nonferrous metals 2
3440 Fabricated structural metal products 1
3452 Bolts, nuts, screws, rivets and washers 1
3510 Engines and turbines 1
3533 Oil and gas field machinery and equipment 1
3540 Metalwork machinery and equipment 1
3541 Machine tools, metal cutting types 1
3559 Special Industry machinery, NEC 8
(continued)
(continued)
SIC Description
3570 Computer and office equipment 1
3571 Electronic equipment 1
3572 Computer storage devices 2
3576 Computer communication equipment 6
3577 Computer peripheral equipment, NEC 8
3578 Calculating and accounting machines (no electronic computers) 2
3620 Electrical industrial apparatus 1
3621 Motors and generators 2
3651 Household audio and video equipment 2
3652 Phonographs records and prerecorded audio tapes and disks 1
3661 Telephone and telegraph apparatus 10
3663 Radio & TV broadcasting and communications equipment 8
3669 Communications equipment, NEC 2
3670 Electronic components and accessories 1
3672 Printed circuit boards 4
3674 Semiconductors and related devices 33
3679 Electronic components, NEC 3
3690 Miscellaneous electrical machinery, equipment and supplies 1
3714 Motor vehicle parts and accessories 2
3716 Motor homes 1
3823 Industrial instruments for measurement, display, and control 3
3825 Instruments for measuring and testing of electricity and electrical signals 5
3826 Laboratory analytical instruments 8
3827 Optical instruments and lenses 3
3829 Measuring and controlling devices, NEC 5
3841 Surgical and medical instruments and apparatus 3
3842 Orthopedic, prosthetic and surgical appliances and supplies 5
3844 X-Ray apparatus and tubes and related irradiation apparatus 2
3845 Electromedical and electrotherapeutics apparatus 10
3861 Photographic equipment and supplies 3
3873 Watches, clocks, clockwork operated devices/parts 1
3942 Dolls and stuffed toys 1
3944 Games, toys and children’s vehicles (no dolls and bicycles) 1
3949 Sporting and athletic goods, NEC 2
3990 Miscellaneous manufacturing industries 2
4213 Trucking (no local) 4
4400 Water transportation 1
4512 Air transportation, scheduled 2
4731 Arrangement of transportation of freight and cargo 2
4812 Radiotelephone communications 9
(continued)
Appendix 5.2 Number of Non-Net Firms by SIC 103
(continued)
SIC Description
4813 Telephone communications (no radiotelephone) 18
4832 Radio broadcasting stations 9
4833 Television broadcasting stations 3
4841 Cable and other pay television services 9
4899 Communications services, NEC 6
4924 Natural gas distribution 1
4953 Refuse systems 3
4955 Hazardous waste management 1
5010 Wholesale—motor vehicles and motor vehicle parts and supplies 1
5020 Wholesale—furniture and home furnishings 1
5045 Wholesale—computers and Peripheral Equipment and Software 3
5047 Wholesale—medical, dental and hospital equipment and supplies 1
5063 Wholesale—electrical apparatus and equipment, wiring supplies 1
5065 Wholesale—electronic parts and equipment, NEC 1
5072 Wholesale—hardware 1
5080 Wholesale—machinery, equipment and supplies 2
5122 Wholesale—drugs, proprietaries and druggists’ sundries 2
5171 Wholesale—petroleum bulk stations and terminals 1
5180 Wholesale—beer, wine and distilled alcoholic beverages 1
5190 Wholesale—miscellaneous nondurable goods 1
5311 Retail—department stores 1
5331 Retail—variety stores 1
5500 Retail—auto dealers and gasoline stations 1
5621 Retail—women’s clothing stores 2
5651 Retail—family clothing stores 1
5700 Retail—home furniture, furnishing and equipment stores 1
5731 Retail—radio, tv and consumer electronics stores 1
5735 Retail—record and prerecorded tape stores 1
5812 Retail—eating places 10
5900 Retail—miscellaneous retail 2
5940 Retail—miscellaneous shopping goods stores 1
5944 Retail—jewellery stores 1
5945 Retail—hobby, toy and games shops 1
5961 Retail—catalog and mail-order houses 4
5990 Retail—retail stores, NEC 1
6020 National Commercial Banks 25
6035 Savings institutions, federally chartered 6
6099 Functions related to depositary banking, NEC 1
6141 Personal credit institutions 2
6153 Short-term business credit institutions 2
(continued)
(continued)
SIC Description
6162 Mortgage bankers and loan correspondents 1
6163 Loan brokers 1
6200 Security and commodity brokers, dealers, exchanges services 1
6211 Security brokers, dealers and flotation companies 1
6282 Investment advice 1
6331 Fire, marine and casualty insurance 1
6351 Surety insurance 1
6411 Insurance agents, brokers and services 4
6513 Operators of apartment buildings 1
6794 Patent owners and lessors 3
6798 Real estate investment trusts 2
7310 Services—advertising 3
7331 Services—direct mail advertising services 1
7359 Services—equipment rental and leasing, NEC 1
7361 Services—employment agencies 2
7363 Services—help supply services 1
7370 Services—computer programming, data processing, etc. 4
7371 Services—computer programming services 2
7372 Services—pre-packaged software 68
7373 Services—computer integrated systems design 14
7374 Services—computer processing and data preparation 1
7380 Services—miscellaneous business services 1
7389 Services—business services, NEC 2
7819 Services—allied to motion picture production 1
7830 Services—motion picture theatres 1
8011 Services—office and clinics of doctors of medicine 1
8071 Services—Medical Laboratories 1
8200 Services—educational services 4
8351 Services—child day care services 1
8700 Services—engineering, accounting, research, management 1
8711 Services—engineering services 2
8731 Services—commercial physical and biological research 13
8734 Services—testing laboratories 1
8741 Services—management services 3
8742 Services—management consulting 4
9995 Non-operating establishments 1
Total 541
Fonte: http://www.sec.gov/info/edgar/siccodes.htm (U.S. Securities Exchange Commission)
References 105
References
Bagnolli M, Kallapur S, Watts SG (2001) Top line and bottom line forecasts: a comparison of
internet during and after the bubble, working paper, Krannert Graduate School of Management,
Purdue University. http://ssrn.com/abstract=274178
income as a function of financial health. J Acc Econ 25:1–34
Bartov E, Mohanram P, Seethamraju C (2002) Valuation of internet stocks—an IPO perspective.
J Acc Res 40(2):321–346
Bowen RM, Davis AK, Rajgopal S (2002) Determinants of Revenues-reporting practices for
internet firms. Contemp Acc Res 19(4):523–562
Burgstahler DC, Dichev I (1997) Earnings adaptation and equity value. Acc Rev 72(2):187–215
values over the past forty years. J Acc Econ 24(1):39–67
Cooper MJ, Dimitrov O, Rau R (2001) A Rose.com by any other name. J Financ 56(6):2371–2388
companies, 3rd edn. Wiley, NY
Compustat Manual (2004) Manual da Base de Dados, Standard & Poor´s Research Insight.
Compustat North America, US
of what has changed? J Acc Econ 34(1&3):43–67
Damodaran A (2001) The dark side of valuation. Prentice Hall, US
Davis AK (2002) The value relevance of revenue for internet firms: does reporting grossed-up or
barter revenue make a difference? J Acc Res 40(2):445–477
Demers E, Lev B (2001) A Rude awakening: the internet shakeout in 2000. Rev Acc Stud 60
(2&3):331–359
Econ 68(3):413–437
DuCharme LL, Rajgopal S, Sefcki SE (2001) Why was internet IPO underpricing so severe?
Working paper, University of Washington. http://ssrn.com/abstract=285986
Eviews Manual (2004) Eviews 5 user’s guide. Quantitative Micro Software. Timberland, UK
Fama E, French K (1998) Taxes, Financing decisions, and firm value. J Financ 53(3):819–842
Francis J, Schipper K (1999) Have financial statements loss their relevance? J Acc Res 37(2):319–
352
Hand JRM (2001a) Evidence on the Winner-Takes-all Business Model: the profitability
returns-to-scale of expenditures on intangibles made by U.S. Internet Firms, 1995-2001,
working paper vol 14. Kenan-Flager Business School, Chapel Hill
Hand JRM (2001b) The role of book income, web traffic, and supply and demand in the pricing of
US Internet stocks. Eur Financ Rev 5(3):295–317
(eds) Intangibles: an oxford management reader. Oxford University Press, Oxford
Hendershott RJ (2001) “Net value: wealth creation (and destruction) during the internet boom”,
working paper. Leavey School of Business, Finance Department, Santa Clara University, Santa
Clara
Joos PR, Plesko GA (2005) Valuing loss firms. Acc Rev 80(3):847–870
Jorion P, Talmor E (2006) Value relevance and non-financial information in emerging industries:
the changing role of web traffic data, working paper, University of California–Irvine e London
Business School. http://ssrn.com/abstract=258869
Kaplan SN (2002) Valuation of new economy firms, working paper. University of Chicago,
Chicago
Keating EK, Lys TZ, Magee RP (2003) Internet downturn: finding valuation factors in spring
2000. J Acc Econ 34(1&3):189–326
VDM
Lee PM (2001) What’s in a name.com?: The effects of “.com” Name changes on stock prices and
trading activity. Strateg Manag J 22:793–804
Lewellen J (2003) Discussion of “The internet downturn: finding valuation factors in spring 2000”.
J Account Econ 1–3:237–247
Liu Q, Song F (2001) The rise and fall of internet stocks: should analysts be blamed? Working
paper. School of Economics and Finance, University of Hong Kong, Hong Kong
Ljungqvist A, Wilhelm WJ (2003) IPO pricing in the Dot.Com bubble. J Financ 58(2):723–752
Loughran T, Ritter J (2003) Why has IPO underpricing changed over time. Working paper
presented at AFS—American Finance Association 2003, Washington Meeting
Loughran T, Ritter J (1995) The new issue puzzle. J Financ 50(1):23–51
pricing, working paper, Polytechnic University of Valência—Department of Economics and
Morgan & Stanley (1999)The internet company handbook. June 1999
Morgan & Stanley (2000) The B2B internet report. April 2000
Morgan & Stanley (2001) The technology IPO yearbook, 7th edn
Morgan & Stanley (2002) The technology IPO yearbook, 8th edn
Neves JC (2002) Avaliação de Empresas e Negócios. McGraw-Hill, Portugal
Noe, T, Parker, G. (2005). Winner Take all: Competition, Strategy and the Structure of Returns in
the Internet Economy, J Econ Manage Strat 14(1):141–164.
Ofek E, Richardson M (2002) The valuation and market rationality of internet stock prices. Oxford
Rev Econ Policy 18(3):265–287
Ofek E, Richardson M (2003) Dot.com Mania: the rise and fall of internet stocks prices. J Financ
58(3):1113–1138
Penman SH (2001) Discussion of “Basic to basics: forecasting the revenues of internet firms” and
“A rude awakening: internet shakeout in 2000”. Rev Acc Stud 2&3:361–364
Penman SH (2003) Financial Statement analysis and security valuation, 2nd edn. McGraw-Hill, US
Ritter J (1991) The long-run performance of initial public offerings. J Financ 46(1):3–27
Rajgopal S, Venkatachalam M, Kotha S (2002) Managerial Actions, stock returns, and earnings:
the case of business-to-business internet firms. J Acc Res 4(2):529–556
Rajgopal S, Venkatachalam M, Kotha S (2003) The value relevance of network advantages: the
case of E-Commerce firms. J Acc Res 41(1):135–162
Schultz P, Zaman M (2001) Do the individuals close to internet firms believe they are overvalued.
J Financ Econ 59(3):347–381
Talmor E (2001) Comment on “The role of book income, web traffic, and supply and demand in
the pricing of U.S. internet stocks. Eur Financ Rev 5:319–321
Tan C (2004) Alternative adaptation of distressed firm’s resources: the valuation roles of equity
and earnings, working paper, New York University. http://ssrn.com/abstract=293125
Tokic D (2004) R&D, advertising and the market value of internet firms: Part I. J Internet Commer
3(2):79–99. doi:10.1300/J179v03n02_05
Tokic D (2005) R&D, advertising and the market value of internet firms: Part II. J Internet
Commer 4(4):23–40. doi:10.1300/J179v04n04_02
Trueman B, Wong MHF, Zhang X (2001) Back to basics: forecasting the revenues of internet
firms. Rev Acc Stud 6:305–329
Chapter 6
Method
Abstract This chapter analyses the phenomenon of “positive valuation of losses”

in companies of the new economy in the US. We start by analysing the life-cycle
effect of those companies. Because the value of a company is a function of the value
of the current assets and the net present value of future growth opportunities, we
examine the pattern of the evolution of the variables investments in total assets,
R&D and advertising over sales and/or debt in the short and medium to long term.
We also include in the analysis the ratio market to book value (MVE/BVE). Based
on the trend analysis, we conclude that the main source of value of net firms and the
non-net firms (match sample), still belong to the future growth opportunities held by
these companies.
Keywords Positive valuation of losses Investment in R&D and advertising

Life cycle effect Research hypotheses
6.1 Introduction
Given the heterogeneous nature of net firms (e.g. Morgan Stanley subdivides this
sector into 11 subsectors), first we split the sample into two group: firms that report
profit and firms that report losses, given the asymmetrical nature with which this
variable is assessed by the market. Then, we proceeded with the partition of the
sample into two subgroups: “B2B—Business-to-Business”—assuming that this
group invests mainly in R&D, and “B2C—Business-to-Consumer”—where now
the predominant investment focuses on advertising. Thus, the empirical analysis
examines four groups of companies: “R&D_B2B with profits and losses” and
“Pub_B2C with profits and losses”.
In addition, because we aim to analyse the relationship between the market
capitalization of net firms and the losses reported by these companies, in point
Sect. 6.4, we begin by respecifying the OM and FOM. Subsequently, we used the

DOI 10.1007/978-981-10-3009-3_6
108 6 Method
methodology of Fama and MacBeth (1973) to analyse how the market evaluates the
net firms based on the main determinants of their value over time (life cycle).
Finally, we formulate the research hypotheses, highlighting the importance of
establishing a comparative analysis between two samples, i.e. net and non-net firms,
in order to ensure robustness of the results (e.g. Lee 2001; Cooper et al. 2001).
6.2 The Effect of Life Cycle
The value of a company is the sum of the value of its total assets from which it
expects to generate revenues in the future and the present value of growth oppor-
tunities (Miller and Modigliani 1961; Feltham and Ohlson 1995; Ohlson 1995;
Copeland et al. 2000; Damodaran 2001; Brealey and Myers 2003). As discussed
earlier, the technology-based companies, as net firms, initially report a high level of
losses as a strategy for maximizing growth. Technology-based companies invest
massively in intangible assets which cannot be capitalized under GAAP. The effect
of the network boosted by the Internet is touted as the main driver of this strategy,
which popularized the concept of winner-takes-all with the aim of quickly
achieving the highest market shares (e.g. Noe and Park 2006).
Thus, given the emerging nature of the Internet sector, we begin by analysing the
evolution of the variables “sales”, results” and “investment” in these
companies/sector. This comparison aims to examine whether the impact of certain
variables on the financial value of the company changes throughout the life cycle of
the company/sector. Thus, we analyse the pattern of evolution of the variables:
“age”, “sales”, “total assets (TA)”, “earnings before extraordinary items
(Res_IExt)”, “total liabilities”, “long-term debt”, “R&D”, “advertising (adv)”,
“equity” and the ratio “MVE/BVE”, “total liabilities/total assets”, “R&D/sales”,
and “advertising/sales”. We examine the “results before items extraordinary
(Res_IExt)” because, given the principle of continuity, it is not sustainable over
time to maintain the effect of extraordinary items (Penman 2003). In discussing the
results we will return to this issue.
From an econometric point of view:
(i) Given the strong asymmetry that characterizes the distribution of the variables
under analysis, we analyse the trend using the average value of the variables,
and also the median value.
(ii) To calculate the ratios of the variables, we followed the method suggested by
Chan et al. (2001), i.e. we aggregate the value of the numerator as well as the
values of the denominator in order to reduce the effect of outliers.
In summary, the trend analysis will be conducted based on the scheme of
Fig. 6.1.
As a first step, we analyse the behaviour of variables over time, i.e. from 1996 to
2003. Since the variables do not reveal a statistically significant trend, we
6.2 The Effect of Life Cycle 109
Trend Analysis: ŷ (Mean / median) = α + β t (t-time)
Yes
βˆ is statistically
significant?
No
There are
structural breaks in Introduce dummy
the relationship? variable – Crash
(Chow test) Yes
No
Analysis of the stability parameter(s) based on the test CUSUM test
Fig. 6.1 Trend analysis
investigate the possibility of structural breaks. To detect the existence of potential

structural breaks, we used the Chow Breakpoint test, which allows a comparison to
be made of the estimated coefficients for two or more subgroups of the sample as
well as checking for any statistically significant differences.1 The criteria for the
subdivision of the sample in different subsets are the responsibility of the investi-
gator (Eviews Manual 2004: 568).
Given the period under study, our criterion for the breakdown of the sample was
the year 2000, the year of the crash. In the presence of structural breaks, we
introduced a dummy variable—“crash”—which takes the value one for the post-
crash period (2000–2003) and zero for the previous period, i.e. 1996–1999.2 We
conclude the trend analysis using the CUSUM test, which is calculated based on
recursive estimation for analysis of the stability of the parameters.3 Verification of
this stability (instability) is relevant based on the results obtained by Demers and
1
See Eviews Manual (2004: 568) for more information about this test.
2
The Hensen test (Johnston and Dinardo 2001: 130) can overcome this arbitrariness of the criteria
for the subdivision of the sample. However, the same is not available in version 5.0 of the EViews.
3
For more detail information see Johnston and Dinardo (2001: 130).
110 6 Method
Lev (2001), Tockic (2005) and Keating et al. (2003). These authors detect changes,
for example for the variables “cash flow”, “results (losses)” and “volatility”, which
lost statistical significance in the post-crash period. Table 6.1 presents the results
for the trend analysis.
The values obtained for trend analysis in the group of net firms (Table 6.1)
shows that:
(i) The variable “age” shows an increase in the mean and median values sig-
nificant at the 1 and 5% level, respectively. The decrease for the year 1999
reflects that the “boom” of the IPOs occurred in this year;
(ii) The variable “sales” indicates a similar pattern, reporting the median value a
statistical significance at the 1% level;
(iii) The variable sales growth reports an increase as well as the investment in the
R&D variable, even after the crash, as well as with the increase of the age of
the company4;
(iv) The variable “Total Assets” has also seen an upward trend, but with a lower
statistical significance (the median value is only statistically significant at the
10% level), which may reflect the predominance of intangible assets (R&D
and advertising investments), which cannot be capitalized under GAAP;
(v) Despite the sales increase, the “results” remain negative (losses). For the
years 2000 and 2001, the periods after the effect of the dot.com bubble and
based on the CUSUM test, the parameter b ^ reveals a strong instability (see
Appendix 6.1). Although having detecting a structural break with the CHOW
test, we found that, with reference to the average and during the crash, the
negative coefficient is statistically significant (significant at the 10% level),
confirming the persistence of the phenomenon of positive valuation of losses.
This result suggests, and in line with the results obtained by McCallig (2004)
and Joos and Plesko (2005), a tendency for high-tech companies to report
losses for longer periods.5
(vi) Regarding the ratio “MVE/BVE”, the coefficient only becomes statistically
significant after the introduction of the variable “crash” (the Chow test
reveals a structural break for the year 2000). As expected, the post-crash
reverses the sign of this coefficient, reinforcing its significance, reflecting the
sharp drop recorded by these companies in terms of market capitalization.
What is surprising, and based on the result of the CUSUM test, is that the
coefficient b ^ proved to be stable over the entire period (see Appendix 6.2);
4
It is possible that the values of this variable are undervalued because in the absence of infor-
mation; for example for the variable R&D (na—not available), we assign a null value, in order to
preserve the sample size.
5
This persistence of the losses, and as noted by Keating et al. (2003), was not due to the dis-
semination of new information to the market after the crash of dot.com bubble, e.g. more pes-
simistic information about the analysts forecasts or even for the performance of the “web traffic”
variables (e.g., number of visitors, pages viewed, time spent in the query data website, and the
growth rate of the Internet-user population).
Table 6.1 Trend analysis for the sample of net firms
Variáblesa 1996 1997 1998 1999 2000 2001 2002 2003 ^
b CUSUM Crashd CUSUM
testc ^pre-crash ^post-crash testc
b b
Ageb Average 1.65 2.06 2.40 1.81 2.28 3.18 4.19 5.29 0.465*** E
Median 1.00 2.00 2 1.00 2.00 3.00 4.00 5.00 0.50** E
Sales Average 111.16 104.90 102.82 87.65 133.20 144.86 153.77 180.22 10.71** E
Median 26.76 35.25 38.09 26.20 46.87 52.04 56.52 65.50 5.24*** E
Total Average 118.42 143.51 161.67 299.25 551.33 342.67 352.20 408.18 46.03** E
Assets Median 52.90 57.54 60.13 86.27 127.82 91.39 95.55 90.76 7.03* E
Results Average −2.37 −6.33 −15.04 −31.07 −111.20 −153.85 −80.56 −18.44 −11.67 I (00 and 12.84 −128.66* E
(Res_IExt) 01)
Median −0.83 −3.07 −5.53 −13.65 −28.76 −34.91 −15.23 −4.73 −2.28 I (00 and 2.54 −25.31 E
01)
Total Average 62.57 79.64 89.61 136.31 211.45 171.32 195.27 245.97 25.98*** E
liabilities Median 9.88 12.56 14.20 16.04 25.37 24.33 26.08 28.36 2.82*** E
Long term Average 20.42 30.26 35.75 58.29 73.55 51.44 43.18 42.08 3.32 E 0.82 13.10 E
debt Median 0.16 0.21 12.10 12.36 00.29 0.13 0.05 0.05 −0.02 E −0.015 −0.017 E
R&D Average 4.57 6.25 7.92 7.20 16.30 20.12 18.38 17.65 2.36** E
Median 3.08 1.15 1.13 2.34 6.43 7.65 6.05 5.20 0.75** E
Advertising Average 0.45 1.37 2.28 4.32 7.29 5.29 4.34 4.50 0.66* E
Median 0.00 0.00 0.00 0.00 0.33 0.26 0.06 0.00 0.016 E −0.06* 00.40** E
BVE Average 55.22 61.46 70.24 151.61 326.75 168.22 153.76 159.12 19.74 E −10.97 161.21 E
Median 31.22 35.56 37.75 60.41 90.47 58.73 47.49 47.70 3.19 E −2.49 29.82 E
MVE/BVE Average 5.54 6.33 8.08 16.70 2.21 2.64 2.18 5.01 −0.66 E 2.16* −14.78** E
Median 4.64 4.62 5.32 11.39 1.25 1.47 1.42 3.40 −0.55 E 1.37* −10.07** E
T.Liab./T.A Average 12.53 0.55 0.55 0.46 0.38 0.50 0.55 0.60 0.003 E 0.025 −0.112 E
Median 0.19 0.22 0.24 0.19 0.20 0.27 0.27 0.32 0.015** E
(continued)
111
112
Variáblesa 1996 1997 1998 1999 2000 2001 2002 2003 ^

b b
R&D/Sales Average 0.04 0.06 0.08 0.08 0.12 0.14 0.12 0.10 0.011** E E
Median 0.12 0.03 0.03 0.10 0.14 0.15 0.11 0.08 0.006 E −0.015 0.11**
Adv/Sales Average 0.004 0.013 0.0222 0.0493 0.0548 0.0365 0.028 0.025 0.003 E 0.002 0.004 E
Median 0.00 0.00 0.00 0.00 0.007 0.005 0.001 0.000 0.003 E −0.001 0.008 E
a
See definition of variables in Table 5.8
b
This variable is calculated with reference to the date of the IPO, because we do not have information at the date of the foundation of the company
c
For the representation of the CUSUM test we adopt the following legend: “E” means stability in behaviour of the parameters, i.e. parameters were found to be in the confidence interval defined
by default (a = 5%) by EViews, “I” reports the opposite scenario. In brackets we indicate whether the last two digits of (s) year (s) in which there was instability in the parameter estimates (b).
For example, we report the result of this test for the variable “results (Res_IExt)” [I] and “MVE/BVE [E] in Appendices 6.1 and 6.2, respectively
d
Estimated parameters after introducing the variable dummy “crash”, which takes the value 1 for the period post-crash (2000–2003) and 0 for the previous period (1996–1999). This variable is
included whenever the Chow test becomes statistically significant
(***), (**) and (*) indicates statistically significant at 1, 5 and 10% level, respectively
6 Method
6.2 The Effect of Life Cycle 113
(vii) It should be noted that the variable “BVE” has a positive coefficient, but not
statistically significant, even after the introduction of the variable “crash”.
The highest value obtained for the year 2000 is easily explained, as it reflects
the capital rose in the IPOs processes.
With no increase in equity, and systematically reporting losses, companies can
only finance investments through the use of debt. Both the “total liabilities” and the
ratio “total liabilities over total assets—T. Liab/T.A.” report a statistically signifi-
cant increasing trend. However, the medium- and long-term increase in a small
magnitude (the variable long-term debt is not statistically significant), thus, the
predominance of short-term debt reflects the high agency costs associated with this
type of company. These are companies operating in an emerging sector, with some
technological complexity and with a very limited historical record. Lenders come in
renegotiating the ceiling of short-term debt, an efficient way to monitor this type of
company (Jensen and Meckling 1976).
Given these results, and with reference to the characterization of the life cycle of
the company/sector proposed by Damodaran (2001: 13), we conclude that in the
period under review, the Internet was still a fast-growing sector. The “sales” con-
tinue to increase in line with investments in “R&D” (see the very similar values for
the mean and median of the ratio “R&D/Sales” [(0.10)(0.08) − 2003], particularly
in the period post-crash. In this context, the persistence of losses as sustained by the
FOM is (in part) a consequence of the conservatism accounting effect, particularly
in start-up companies. Strong pressures at the level of liquidity, which such com-
panies were subjected to after the crash, and in line with the results obtained by
Demers and Lev (2001) and Keating et al. (2003), are overcome using a strategy of
merger and acquisitions (M&A), as documented in Sect. 5.4 (see Tables 5.3 and 5.
4). The M&A strategy enables companies not only to overcome the situation of
financial stress, but also to achieve the desired dimension confirming the winner-
takes-all strategy (Noe and Parker 2005). The fact that the processes of M&A have
occurred mainly through the exchange of shares, according to data from Schultz
and Zaman (2001), suggests a clear sign of confidence of both parties (buyer and
the acquired companies) in this sector.
Therefore, we emphasize that the main source of value of these companies is still
associated with the exercise of future growth opportunities. The recovery of the
ratio MVE/BVE in the years 2002 and 2003 [the mean and median values for these
years are 2.18 (1.42) and 5.01 (3.40), respectively], is justified in our view by the
fact that, after the euphoria generated about this sector, and due to consecutive
investments in R&D, the expectations of future cash flows of this sector are still
viewed as positive signs by the market. Table 6.2 shows the results for the sample
non-net firms.
The results for the sample of non-net firms (Table 6.2) are similar to those
obtained for the sample of net firms, revealing, however, a greater statistical sig-
nificance. Thus, the trend of increase in “sales” is accompanied by an increase in
Total Assets, the investment in “R&D” and “advertising”. The “results” also
negative are persistent throughout the period under review (1996–2003),
Table 6.2 Trend analysis for the sample of non-net firms
114
Variblesa 1996 1997 1998 1999 2000 2001 2002 2003 ^

b b
Ageb Average 1.99 2.38 2.98 2.59 2.87 3.86 4.89 5.21 0.453*** E
Median 1.00 2.00 3.00 2.00 2.00 3.00 4.00 4.00 0.369*** E
Sales Average 92.47 109.94 131.44 131.53 153.32 175.67 199.78 237.63 19.28*** E
***
Median 41.41 44.36 60.31 55.57 56.47 64.23 67.97 82.42 4.974 E
Total Assets Average 111.38 130.05 199.55 291.07 356.36 397.74 409.51 466.69 54.01*** E
Median 49.51 59.84 81.89 93.63 122.87 137.20 146.69 159.55 16.66*** E
Results Average 1.08 −0.86 −0.24 −6.10 −22.84 −37.11 −33.88 −10.85 −4.476** I (02)
(Res_IExt) Median 0.92 0.85 1.50 0.25 −2.49 −5.67 −3.04 −1.69 −0.738** E
Total liabilities Average 58.15 72.72 127.88 187.58 208.77 235.91 261.85 306.06 36.027*** E
Median 12.38 17.73 25.48 27.88 29.27 31.15 34.12 41.83 3.65*** E
LongTermDebt Average 21.85 23.52 36.77 88.00 109.67 132.32 152.95 177.12 24.314*** I (99 and
00)
Median 12:56 0.97 1.45 1.76 1.18 0.98 1.41 0.91 0.032 E −0.811 −1.56 E
R&D Average 5.05 5.48 6.86 6.28 10.28 15.57 16.20 17.66 2.048*** E
Median 12:51 0.96 0004 0.00 2.44 4.81 5.20 4.01 0.745** E
Advertising Average 0.48 0.43 0.64 0.91 1.14 1.1 1.40 17.58 1.503* I
(00,01,02)
Median 0.00 0.00 0.00 0.00 0.00 20.00 0.00 0.00 −e −
BVE Average 52.47 56.11 69.59 96.61 136.76 156.49 142.20 156.89 17.41*** E
Median 33.78 37.43 44.35 60.66 73.20 73.18 66.84 74.39 6.43*** E
MVE/BVE Average 3.40 4.47 4.47 6.40 3.41 2.77 1.77 3.00 −0.291 E 0.34 −3.31**
Median 2.77 2.76 1.99 2.95 2.04 2.05 1.35 2.71 −0.097f E
Liab./TA Average 0.52 0.56 0.64 0.59 0.64 0.66 0.014** E
Median 0.25 0.31 0.24 0.23 0.23 0.26 −0.006 E 0.012* −0.096* E
(continued)
6 Method
Variblesa 1996 1997 1998 1999 2000 2001 2002 2003 ^
b b
R&D/Sales Average 0.055 0.05 0.0522 0.0478 0.067 0.0886 0.0811 0.0743 0.005** E
Median 0.012 0.022 0.0001 0.000 0.0431 0.0749 0.0765 0.0487 0.009** E
Adv/Sales Average 0.005 0.004 0.005 0.007 0.007 0.006 0.007 0.074 0.006 I(00,01,02) 0.01 −0.003 R (02)
Median 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 −e −e
a
See definition of variables in Table 5.8
b
This variable is calculated with reference to the date of the IPO, because we do not have information at the date of the foundation of the company
c
For the representation of CUSUM test, we adopt the following description: “E” stands for stability in the behaviour of the parameters, i.e. the parameters belong in the confidence interval
defined by default (a = 5%) by EViews; “I” reports the opposite scenario. In brackets, we indicate the last two digits of (s) year (s) in which instability occurred. As with the net firms, we
present the output for this test for the variables “results (I)” (Appendix 6.3) and the ratio “MVE/BVE (E)” (Appendix 6.4)
d
Estimated parameters after introducing the variable dummy “crash”, which takes the value 1 for the period post-crash 2000–2003 and 0 for the previous period 1996–1999. This variable is
included whenever the Chow test becomes statistically significant
e
Values are always zero
f
The variable “crash” was not introduced because no structural break was detected
***
( ), (**) and (*) indicates statistically significant at 1, 5 and 10% level, respectively
115
116 6 Method
confirming once again the phenomenon of positive valuation of losses. However,

the magnitude of their values is much lower than the values reported by net firms
(see Appendix 6.3). The trend in the ratio “MVE/BVE” and in line with the group
of net firms is negative and not statistically significant. Only with the introduction
of the variable “crash” (due to the breakdown of the structure verified by the Chow
test and only for the mean), does the coefficient for this variable assume negative
values statistically significant post-crash (see Appendix 6.4).
As in the net firms, and unsurprisingly given the prevalence of high-tech com-
panies (see Table 5.6), this group also experiences strong growth, where the main
source of value is associated with future growth opportunities, based on the values
reported by the variables R&D and R&D/Sales. However, growth appears to be
more sustained, because the funding is allocated by indebtedness, in the long term.
In addition, the increase in the variable BVE showed a statistical significance of 1%
for both the average value and the median values for the 8 years under review. In
the group of net firms, this variable was not proven to be statistically significant, nor
was it for the mean values or the median values.
6.3 Criteria for Subdivision of Samples
From the trend analysis, we conclude that the Internet sector is still growing fast.
The main source of value for net firms is the future growth opportunities. The
successive investments in intangible assets (R&D and advertising), and assuming
the principle of rationality, are positive signals to the market about the future value
of cash flows. The current losses appear to be the consequence of the effect of
conservatism accounting, as modelled by the FOM. A similar pattern of behaviour
is recorded for the sample of non-net firms, an expected result, given the prevalence
also in this sample of high-tech companies. But to make a better analysis of this
effect, i.e. conservatism accounting, it is important to split the group of firms into
those firms that report profits and those losses.6
According to the theory of abandonment option (Hayn 1995; Chambers 1997;
Subranmanyam and Wild 1996), losses assume a transitory nature, so they are
irrelevant for estimating future cash flows, and consequently, the value of a com-
pany. The persistence of losses makes the abandonment option more valuable for
shareholders because it gives them the option to liquidate the company. But given
the company’s profile in the samples under study, high-tech companies in the start-
up phase, and given the type of investment that tends to characterize those firms,
losses are not an adequate proxy for the likelihood of the abandonment option for
6
For example, Trueman et al. (2000: 151) recognized that given the limited number of observa-
tions (217 observations for the quarters September 1998 to December 1999), they were unable to
estimate the model of OM expanded with the introduction of the web traffic variables for the two
groups of companies: companies with profits and companies with losses.
6.3 Criteria for Subdivision of Samples 117
these companies. In this context, the variable results assume importance as a proxy
for the phase of the life cycle of the company.7
The choice of this variable as a proxy for the stage of the life cycle is based on
the duality of the effect of the short-, medium- and long-term association with
investments in intangible assets (Sougiannis 1994). In the short term, given the
imposition of GAAP, these investments are treated as costs. Thus, only in the
medium and long term [5–9 years, according to Lev and Sougiannis (1996)] it
is expected that these companies will start to generate positive cash flows.
Thus, it is assumed that a company with profits has already reached a stage of
stable/mature growth, while a company with losses is identified as a company under
start-up with accelerated growth.8 This classification captures four effects: (i) un-
recognized assets associated with intangible assets (R&D and adv); (ii) differences
in persistence in the growth of various items along the life cycle of the company;
(iii) the impact of risk; and (iv) the size effect (Barth et al. 1998).
To better characterize the two groups of companies and to examine the accuracy
of the choice of the variable “results” as a proxy of the stage of the life cycle of
companies, we analysed in Tables 6.3 and 6.4, with reference to both samples, i.e.
net firms and non-net firms, the statistically significant differences between means
and medians for the variables: “age”, “MVE”, BVE”, “results”, “sales”, “R&D”,
“advertising (adv)”, “R&D/sales” and “advertising(adv)/sales” by year (i.e. 1996–
2003).
Based on the results shown in Table 6.3 (panel A and B), the results of the
variable MVE are statistically different at 1% for the median values, except for the
years 1998 and 1999. This result is easy to justify, since the year 1999 was the year
that the net firms approached the peak of the market capitalization (see Fig. 6.2).
Note that the value of the variable MVE is systematically higher for the group of
companies that have profits compared to the group with losses, with the exception
of 1998, and with reference to the average group whose difference is not statistically
significant.
For the variable BVE, the results for the mean values differ substantially from
the median values. Given the strong asymmetry that characterizes the distribution of
the variables under analysis, the study focuses on the results for the median. Thus,
for all the years, except 1998, the differences are statistically significant; moreover,
the corresponding value of equity for the group with losses is about the half of the
value reported by the group with profits (48.63 million compared to 81.63 million
US dollars—median values).
As expected, the losses are systematic throughout the period for the group that
recorded losses, which is also the largest group of companies. This persistence is
7
The variable “age” is not appropriate, because it is calculated with reference to the date of the IPO
of the company.
8
McCallig (2004) complements the analysis using the analysis of the variable “retained earnings”.
However, this author selects all listed companies in the US markets (NASDAQ, NYSE, and
AMEX) since 1980 and does not impose any criteria for selecting the sample. Thus, the companies
selected are at different stages of the life cycle.
Table 6.3 Differences between means/medians in the sample of net firms: companies with profits and companies with losses
118
Variablesb Nc Age MVE BVE (Res_IExt) Sales R&D Adv R&D/Sales Adv/Sales
Periods
Panel A: Meansa
1996 Profits 34 1.75 480.5** 86.2* 11.09** 205.6** 5.06 0.17 0.075 0.004*
Losses 37 1.53 145.8 26.8 14.7 24.36 4.12 0.70 4.402 0.052
1997 Profits 49 2.23 583.9* 89.9* 12.98** 183.9** 7.79 1.31 0.069 0.008*
Losses 66 1.95 244.9 40.4 20.7 46.24 5.10 1.41 1.223 0.063
1998 Profits 47 2.27 469.4 90.9 12.77** 194.1** 8.69 2.56 0.065 0.016*
Losses 107 2.72 610.8 81.2 27.3 62.75 7.58 2.16 0.671 0.097
1999 Profits 80 2.61 4055.5* 181.6 18.35** 210.8** 11.6** 5.12 0.07 0.025
Losses 350 1.63 2184.1 144.8 42.4 59.5 6.19 4.14 0.798 0.298
2000 Profits 74 3.07 1207.5* 259.5 22.6** 279.6** 12.6 9.95 0.09** 0.026*
Losses 453 2.13 644.4 337.7 133.1 109.3 16.9 6.93 0.37 0.204
2001 Profits 36 4.06 1207.1** 255.6 20.02** 267.7** 14.2 7.35 0.081 0.022
Losses 392 3.10 374.6 160.2 169.8 133.6 20.7 5.10 0.439 0.064
2002 Profits 54 4.49 1001.2** 288.9** 24.26** 256.7* 14.3 10.6** 0.082* 0.026
Losses 313 4.14 220.2 130.4 98.6 136.0 19.1 3.3 0.387 0.029
2003 Earnings 101 5.49 1798.7** 301.5** 29.30** 315.7** 21.1 9.94** 0.09 0.015
Losses 206 5.19 306.1 89.3 41.8 113.8 15.9 1.83 0.69 0.014
Panel Profits 475 4.28 1586.7** 211.3 20.36** 235.4** 13.0 6.57** 0.093* 0.022**
Datad Losses 1924 4.57 740.64 174.6 96.9 118.1 14.8 4.34 0.633 0.133
Panel B: Mediansd
1996 Profits 34 1 272.05** 59.23** 5.10** 71.24** 4.27 0 0.022** 0
Losses 37 1 105.19 21.02 8.45 10.18 1.96 0 0.234 0
1997 Profits 49 2 229.27** 49.55** 4.92** 80.19** 0 0 0** 0
Losses 66 2 111.54 18.27 10.78 16.9 1.23 0 0.193 0
6 Method
(continued)
Variablesb Nc Age MVE BVE (Res_IExt) Sales R&D Adv R&D/Sales Adv/Sales
Periods
1998 Profits 47 2 251.19 48.87 7.67** 106.1** 0.2 0 0.003** 0
Losses 107 2 168.71 42.47 11.29 22.31 1.13 0 0.123 0
1999 Profits 80 1 825.89 93.10** 7.09** 102.7** 2 0 0.011** 0**
Losses 350 2 678.28 56.0 19.35 21.28 2.34 0.09 0.161 0.005
2000 Profits 74 2 249.93** 116.9 8.83** 129.2** 4.96 0e 0.069** 0**
*
Losses 453 2 100.74 84.95 36.47 39.87 6.63 0.38 0.20 0.014
2001 Profits 36 3 459.78** 119.7* 7.49** 136.3** 5.14 0.03 0.036** 0.001
Losses 392 3 76.69 56.65 40.65 47.96 7.76 0.27 0.198 0.006
2002 Profits 54 4 261.74** 69.77** 10.87** 107.8** 2.79 0.04 0.027** 0.001
Losses 313 4 52.73 41.63 20.13 48.62 6.38 0.06 0.184 0.002
2003 Profits 101 5 461.24** 100.7** 10.32** 104.3** 5.77 0.04 0.08** 0.001
Losses 206 5 107.37 23.92 12.09 43.53 4.73 0 0.182 0
Panel Profits 475 4 368.29** 81.63** 7.42** 98.53** 3.56 0.01** 0.033** 0**
Data Losses 1924 4 128.40 48.63 24.29 33.75 4.89 0.06 0.185 0.002
a
The test of equality of means tests whether each subgroup has the same mean. If so, then the mean variance between groups must be equal to the variance
within each group. As we analyse the equality of means between only two subgroups, the EViews reports the statistic T, which are obtained from the square
root of the F statistic, assuming only one degree of freedom in the numerator
b
Variables are in millions of dollars, except ratios. See definition of variables in Table 5.8
c
N = Number of firms in each group
d
The test for differences between medians is the non-parametric Kurskal–Wallis one-way ANOVA by ranks. This test is a generalizing Mann–Whitney test for
more than two groups. The basic intuition is to establish ranks among various subgroups. If the sum of subgroup 1 is identical to the sum of subgroup 2, then
the two groups have identical values for the median. EViews also provides other non-parametric tests, such as: the Wilcoxon signed ranks test, the Chi-square
test and the van der Waerden (normal scores) test (EViews Manual 2004: 309)
e
The difference is only statistically significant at a significance level of 10%, according to the v2 test
**
( ) and (*) identifies statistical differences significantly at 1 and 5% level, respectively
119
Table 6.4 Differences between means/medians in the sample of non-net firms: companies with profits and companies with losses
120
Variablesb Nc Age MVE BVE (Res_IExt) Sales R&D Adv R&D/sales Adv/sales
Periods
Panel A: Meana
1996 Profits 81 2.17 214.06 59.18 7.36* 114.1* 4.8 0.46 0.07* 0.004*
Losses 62 1.76 131.55 43.70 7.11 64.27 5.36 0.50 0.84 0.03
1997 Profits 109 2.36 307.54 70.05** 8.80** 256.0** 4.9 0.30 0.06 0.003*
Losses 82 2.41 175.83 37.58 13.7 48.67 6.25 0.60 6.51 0.017
1998 Profits 120 3.09 452.59* 92.19** 12.71** 200.5** 5.81 0.82 0.04 0.004
Losses 92 2.83 126.73 40.11 17.13 41.4 8.23 0.41 13.38 0.171
1999 Profits 182 2.72 441.34* 86.50 10.48** 174.7** 4.83* 0.94 0.05 0.005
Losses 174 2.44 803.44 107.2 23.44 86.36 7.8 0.87 1.63 0.40
2000 Profits 195 3.30 494.29 111.5 13.93** 214.8** 6.85** 1.02 0.05** 0.005
Losses 250 2.55 443.74 156.4 51.52 105.4 13.0 1.23 2.2 0.08
2001 Profits 140 4.35 607.92** 146.7 14.26** 243.8** 11.2* 1.51 0.07** 0.006
Losses 262 3.61 339.06 161.7 64.55 139.3 17.9 0.92 1.87 0.012
2002 Profits 145 5.49 378.91** 157.7 16.75** 292.1** 9.81** 1.76 0.06** 0.006
Losses 229 4.52 170.60 132.4 65.94 141.4 20.4 1.18 1.63 0.018
2003 Profits 146 5.18 684.32** 230.2** 23.91** 357.9** 15.0** 38.3 0.06* 0.137
Losses 190 5.24 306.77 100.6 37.56 145.2 19.7 1.64 1.67 0.013
Panel Profits 1118 4.28 466.77** 123.4 13.97** 225.8** 8.16** 5.9 0.06** 0.022
Data Losses 1341 4.98 351.35 118.6 44.18 111.6 14.4 1.04 2.88 0.089
Panel B: Mediansd
1996 Profits 81 2 123.01** 38.5** 4.79** 59.53** 0.51 0 0.018 0
Losses 62 1 66.13 23.88 −3.83 16.28 0.62 0 0.044 0
1997 Profits 109 2 126.59** 45.89** 4.19** 77.24** 0.05* 0 0.0004** 0
Losses 82 2 58.22 23.01 18.33 2.67 0 0.125 0
6 Method
−6.19
(continued)
Variablesb Nc Age MVE BVE (Res_IExt) Sales R&D Adv R&D/sales Adv/sales
Periods
1998 Profits 120 2 102.08** 52.45** 5.71** 94.98** 0** 0 0** 0
Losses 92 3 50.19 25.69 −7.45 27.2 4.31 0 0.14 0
1999 Profits 182 2 142.6 57.41** 5.84** 88.49** 0** 0 0** 0
Losses 174 2 163.75 37.89 −10.08 24.82 2.34 0 0.11 0
2000 Profits 195 2.5 151.76 72.52 7.03** 115.8** 0** 0 0** 0
Losses 250 2 149.13 74.94 −18.61 32.56 7.18 0 0.189 0
2001 Profits 140 4 246.41** 116.7** 8.44** 131.0** 0** 0 0** 0
Losses 262 3 120.25 63.75 −20.36 38.34 8.61 0 0.184 0
2002 Profits 145 5 158.98** 96.97** 7.02** 124.2** 0** 0 0** 0

Losses 229 4 70.92 50.11 −22.26 39.3 10.2 0 0.212 0
2003 Profits 146 5 297.79** 147.2** 9.48** 152.1** 0** 0 0** 0
Losses 190 4 134.46 48.82 −14.9 48.13 8.41 0 0.196 0
Panel Profits 1118 4 158.0** 67.38** 6.32** 103.8** 0** 0 0** 0
Data Losses 1341 4 101.66 46.95 14.67 33.53 5.81 0 0.176 0
a
The test of equality of means tests whether each subgroup has the same mean. If so, then the mean variance between groups must be equal to the variance
within each group. As we analyse the equality of means between only two subgroups, the EViews reports the t statistics, which are obtained from the square
root of the F statistic, assuming only one degree of freedom in the numerator
b
Variables are in millions of dollars, except ratios. See definition of variables in Table 5.8
c
N = Number of firms in each group
d
The test for differences between medians is based on the non-parametric Kurskal–Wallis one-way ANOVA by ranks. This test is a generalizing Mann–Whitney
test for more than two groups. The basic intuition is to establish ranks among various subgroups. If the sum of the subgroup 1 is identical to the sum of
subgroup 2, then the two groups have identical values for the median. The EViews also provides other non-parametric tests, such as the Wilcoxon signed ranks
test, the Chi-square test and the van der Waerden (normal scores) test (EViews Manual 2004: 309)
(**) and (*) identifies statistically significant differences at 1 and 5% level, respectively
121
122 6 Method
Fig. 6.2 Dow Jones Internet

Composite Index—Monthly
Closing Prices in the period
July 1997 to February 2002
Internet Composite Index
not changed even after the crash (i.e. the year 2000) in line with the results that we
have obtained for trend analysis. The variable “sales” for this group recorded
successive increases over time; however, the value corresponded to one-third of the
volume of sales recorded for the group of companies with profits (33.75 million
dollars loss group; 98.53 million dollars in profits group—median values).
Now focusing the analysis on the R&D and advertising variables, there are no
relevant differences in investment volumes in the two groups (i.e. profit and loss
companies). Indeed, the group that reported losses invests approximately mean
values of $14.8 million dollars in R&D and reported sales of 33.75 million dollars
(median values); while the group with profits reported an investment in R&D of
$13 million dollars with sales of 98.53 million dollars. Given that a reduction in
these items (R&D and advertising) would allow a substantial reduction in the
volume of short-term losses, these results indicate that a decrease in the investment
in those variables, particularly in R&D in companies that report losses, could be
seen as a sign of confidence on the part of the managers, as well as by the investors,
regarding the good prospects for future profitability associated with this type of
investment.
This conclusion is reinforced when we analyse the R&D/sales ratio. The dif-
ferences between medians are statistically significant at 1% for all years. We
emphasize that this ratio lies in approximately 20% for the group of companies with
losses (panel data), and that the level tends to remain steady even after the crash,
against 3.3% in the group with profits (also according to panel data).
In summary, the results of the sample net firms reflect a more aggressive policy
of investment in intangible assets by the group of firms with losses, particularly in
R&D (note that the sales volume of this group is about one-third of that of the group
with profits). This policy suggests that the losses (partially) derive from the
6.3 Criteria for Subdivision of Samples 123
conservatism accounting effect. In this context, the splitting of the sample between
firms with profits and companies with losses proves to be a good proxy. Table 6.4
now focuses our analysis for the sample of non-net firms.
By conducting a similar analysis for the sample of non-net firms, i.e. comparing
the results obtained for the two groups of companies (i.e. profit and loss compa-
nies), for the variable MVE, except the period of the crash (1999 and 2000), the
differences between the two groups for the median are statistically significant at 1%
level. However, given the panel data values, we note that the difference between the
two groups for the median is not as sharp (158 million dollars for the group with
profits compared to 101.66 million in the group with losses) as for the sample of net
firms (368.29 million dollars for the group with profits compared to 128.40 million
dollars for the group with losses).
With reference to the variable BVE, the behaviour of this variable is similar to
the group of net firms. With reference to the mean value, the differences are not
significant for 5 years, contrary to the behaviour of the median, where the differ-
ences are significant for the entire period.
The differences for the two groups of companies, by reference to the
average/median values of the variable “results (Res_IExt)”, as expected, are sys-
tematically significant at a 1% level. However, it should be noted that the magni-
tude of the losses is small compared to the values recorded for the group of net firms
with losses. The median value for the group of non-net firms with losses is 14.67
million dollars (panel data) compared with 24.29 million dollars (panel data) for the
sample of net firms with losses.
With regard to the sales variable, the behaviour does not differ between the two
samples and the differences are systematically significant for either the mean and
median values. The median value of this variable (panel data) in the sample of non-
net firms for the group with losses amounts to 33.53 million dollars, a value similar
to that recorded by the net firms group with losses (33.75 million dollars).
Analysing the R&D variable, the differences between the two groups (companies
with profits and companies with losses) are systematically significant at 1% level of
significance. This result differs from that obtained for net firms. However, the group
with losses invests more resources, $5.8 million (median values—panel data),
compared with an approximately zero value (also for the median values) for the
group with profits, a similar behaviour reported by the net firms. This differential
behaviour is reflected in the ratio R&D/sales. Because the analyses base on mean
values should be conducted with caution, given the strong asymmetry in the dis-
tribution of this variable, we focus the analysis on the median values.9 Thus,
similarly to the group with losses in the sample of net firms, the share of revenue
from sales and investment that affects R&D stands at 17.6% (panel data), a figure
that shows an increasing trend, even after the crash.
9
See Appendix 6.5, where we report the number of companies by year, with reference to the group
with losses in both samples, when the value of investment in R&D exceeds the value of sales.
124 6 Method
Table 6.5 Split of the Non-net firms Net firms

samples according to business
R&D Advertising (Adv)
model
Profits R&D_B2B Pub_B2C
Losses R&D_B2B Pub_B2C
In summary, for both samples and their groups (companies with profits and
companies with losses), the variables under review recorded a similar behaviour:
(i) there is clear evidence of a more aggressive investment strategy by the group of
companies making investment in R&D compared with advertising, and (ii) in both
samples and for the group of companies with losses, the share of sales that affects
the investments in R&D is about 20% (median value—panel data). Thus, the results
suggest that the losses tend to persist for longer periods, and are justified, in part,
based on the effect of conservatism accounting. In both samples the group with
losses is the one that registers more observations over the period under review
(1996–2003). This result is consistent with the results obtained by trend analysis
and the results of McCallig (2004) and Joos and Plesko (2005).
Given the heterogeneity that characterizes the business model adopted by net
firms, which can be differentiated as business-to-business (B2B) and business-to-
consumer (B2C), each sample (companies with profits and companies with losses)
is further subdivided into two subgroups based on the values recorded by the R&D
and advertising variables.10 Thus, we assume that the B2B group invests primarily
in R&D, while the B2C group favours investment in advertising. As a criterion, the
study defines: when the variable R&D is reported as greater (lower) than the
variable advertising, the company is classified in group B2B (B2C). For compar-
ative purposes, we extend this partition of the non-net firms sample in Table 6.5.
6.4 Methodology of Fama and MacBeth
Given the persistence and magnitude of the losses recorded in both samples (the
group with losses is consistently the group with more observations—Tables 6.3 and
6.4), our objective is to analyse the (statistical) relationship between market capi-
talization and net results reported by these firms over the new economy period
(NEP).
10
The Morgan Stanley reports (The Technology IPO Yearbook—8th and 7th edition) subdivide the
Internet sector into 11 subsectors: (1) Internet Portal; (2) Internet Commerce; (3) Internet
Infrastructure; (4) Internet B2B Software; (5) Internet Financial Services; (6) Vertical Portal;
(7) Internet Infrastructure Services; (8) Internet Consulting & Application; (9) Internet Advertising
& Direct Marketing Services; (10) B2B Commerce; and (11) Multi-sector Internet Companies.
However, subdividing the sample into these 11 subsectors is not possible from an econometric
point of view, given the small number of observations for certain subsectors.
6.4 Methodology of Fama and MacBeth 125
The empirical model to estimate is the OM. Thus, and rewriting the Eq. 2.8:
Pt ¼ bvt þ a1 xat þ a2 vt
and substituting the variable abnormal results (xat —Eq. 2.4) in the previous
expression, we obtain:

Pt ¼ bvt þ a1 xt Rf 1 bvt1 þ a2 vt ð6:1Þ
Assuming the CSR principle, the previous expression can be rewritten as:

Pt ¼ ðbvt1 þ xt dt Þ þ a1 xt a1 Rf 1 bvt1 þ a2 vt ð6:2Þ
Since the vt variable “other information” is an orthogonal variable, the inde-

pendent variable—xat 11 variable assumes a zero value, and by introducing the
constant, which in part “aggregates” the effects associated with the variable “vt,”
i.e. the “shocks” from other sources of information other than financial information,
which are immediately incorporated into prices and only later are reflected in the
financial statements, we obtain the empirical model to estimate:

ðPt þ dt Þ ¼ ð1 þ a1 Þxt þ 1 a1 ðRf 1Þ bvt1 þ a2 vt ,
ð6:3Þ
MVEit ¼ a0 þ a1 xit þ a2 bvi;t1 þ eit
The equation clearly shows the relevance of variables “net income—xit” and
“equity (lag 1 year)—bvi,t−1” as the main value drivers of the OM (Ohlson 1995:
670).12 Because our main objective is to analyse the statistical relationships
between the accounting and financial variables, like Francis and Schipper (1999:
327), we assume that the intrinsic value of securities and market value have the
same structure.
Assuming MVE as a dependent variable, the model (Eq. 6.3) will be affected by
large companies (scale-effect) (e.g. Easton and Sommer 2003). Accordingly, we
apply the logarithmic transformation to the variable MVE, which can compress the
range of the variation, making more uniform the variance of the error, thereby
controlling more easily the effect of outliers. This procedure was also adopted by
Berger et al. (1996), Francis and Schipper (1999) and Hand (2001), among others.13
The transformed model now corresponds to a log-linear model.14
11
Given the orthogonality of this variable, its omission from the model does not bias the estimation
of the remaining coefficients of the model (Greene 2008).
12
By including in the model the variable equity at the beginning of the year, we exclude the effect
of net income variable, also an independent variable in the model.
13
With this transformation, the distribution of the dependent variable is altered, but the new
distribution indicates strong adherence to the normal distribution assumption of the ordinary least
squares method (OLS) (Gujarati 2002).
14
See Greene (2008) for an interpretation of the coefficients of this model.
126 6 Method
Since the main objective is to analyse how the MVE varies over time, i.e. the life
cycle, the main factors identified by the OM model (“net income” and “equity”) as
the main value drivers and in line with the methodology of Fama and MacBeth
(1973), we estimate a regression for each year separately applying the ordinary least
squares method (OLS). Thus, our inference will be based on the average of the
parameters estimated for the period 1996–2003 (i.e. NEP).
Analytically, the model (Eq. 6.3) takes the following form:
X
H
MVEit ¼ a0 þ aj;t Fi;j;t þ eit com i ¼ 1; 2. . .N ð6:4Þ
j¼1
where H identifies the number of explanatory variables included in the model, N the
number of listed companies in the sample, and Fi,j,t represent the explanatory
variable j for firm i at time t (1996–2003). The null hypothesis to test is:
PT
ajt
H0 ¼ t¼1
¼ 0 for t ¼ 1; 2; . . .T
T
i.e. the average of the time series of the estimated parameters over time is zero.
The main advantages of this methodology according to Fama and French (1998)
are as follows: (i) greater control of the survival effect (survivor bias effect), because
it does not require companies to have a long lifetime; this effect is relevant in our
samples (see Tables 5.3 and 5.4, when we analyse the “movements” of entries and
exit in the market by these companies—net and non-net firms); (ii) because the
regression is based on cross-section data, each year we can always include a higher
number of observations compared to the time series; (iii) as it is possible to include
a larger number of observations over the year, this increases the precision of esti-
mated parameters; and (iv) making inferences based on the average value of the
parameter(s) estimated reduces the effect of volatility recorded from year to year,
which is particularly relevant in the present investigation, since the effect of the
dot.com bubble affected the Internet sector from 1999 to 2000 (see Fig. 6.2).
Next we formulate the research hypotheses to be tested.
6.5 Research Hypotheses
Empirical research on losses is scarce, and the results obtained are so far contra-
dictory. Traditionally the information content of losses for evaluation was very low,
given their transitory nature. In this context, the shareholders could exercise the
abandonment option which they hold on firms (Hayn 1995; Chambers 1997;
Subramanyam and Wild 1996).
However, post 90s, an increase in the number of listed companies that report
losses can be observed, including many non-net firms (Hayn 1995; Collins et al.
6.5 Research Hypotheses 127
1997, 1999; McCallig 2004; Joos and Plesko 2005). Hayn (1995) had noted that
those firms tend to be smaller and operate in high-tech sectors. As a result, these
companies tend to favour investments in intangible assets, particularly in R&D and
advertising items (Chan et al. 2001; McCallig 2004; Joos and Plesko 2005).
As modelled by the FOM model, high investments in intangible assets and the
effect of accounting conservatism justify that companies, particularly
technology-based ones and companies in the start-up phase register large losses
because a significant portion of their investment is considered costs (FOM—
Proposition 9). However, due to the magnitude and profile of the investments as in
intangible assets, there are great expectations about future abnormal returns, which
in turn sustain the high market capitalizations which these companies report. In this
context arises the question about the adequacy of losses, or rather their persistence,
as a proxy to pursue the abandonment option in this group of companies.
Thus, given:
(i) The effect of conservatism accounting modelled by Feltham and Ohlson
(1995), which reflects the understatement of the variable “operating assets”
given the accounting costs as investments in R&D and advertising, which are
prevalent in technology-based companies in the start-up phase;
(ii) The definition of the variable “operating cash flows”, and the understatement
of the variable “operating assets—oat” that has to be compensated for by an
overestimation of “abnormal future operating results—oxat ” (FOM): Hence,
the variable “operating cash flows” remains unchanged, given the indepen-
dence of any accounting policies15;
(iii) The investment in R&D and advertising tends to predominate in the samples
under study, which increases with sales and age (Tables 6.1 and 6.2); and
(iv) The fact that this investment strategy tends to be more aggressive in the group
of companies that recorded losses, since for this group the ratio R&D/sales
assumes higher values, which in our view is a clear sign of the confidence of
managers in terms of future profitability (lower investment in such assets
allows the company to decrease in the short-term the volume of losses),
we formulate the first research hypothesis:
Hypothesis 1: For younger firms (net and non-net firms) losses are positively
valued by the market
With the formulation of the first hypothesis, and in line with the results of
McCallig (2004) and Joos and Plesko (2005), we assume—outside the universe of
net firms—the statistical significance of the phenomenon of positive valuation of
losses of technology-based companies in the start-up/growth phase by the market
(this criteria agrees with the companies under study).
P1
Remember that the operating cash flow variable is defined as:
15
Rs
f Et ðct þ s Þ ¼
P a s¼1
oat þ 1 s
s¼1 Rf Et oxt þ s (se footnote 15).
128 6 Method
To empirically test this hypothesis, we express the market value of the company
(MVE) equity on the results before extraordinary items.16 The model to be tested is
defined as follows:
MVEit ¼ a0 þ a1 ðRes IExtÞit þ eit ð6:5Þ
with i = 1, 2 … N and t = 1996–2003.

The phenomenon—positive valuation of losses—is a new phenomenon, and the
results so far have achieved little consensus. Collins et al. (1999) argue that this
phenomenon is due to an incorrect specification of the evaluation model. Thus, they
criticize the valuation models based on the capitalization of results (earnings model)
used by Hayn (1995). They empirically show the relevance of the inclusion of the
variable BVE in the evaluation model, arguing that the negative coefficient obtained
for the variable results, is due to the omission of the variable BVE of the evaluation
OM and FOM models.
The theoretical explanation seems to be consensual on the relevance of the
variable BVE for evaluation purposes, despite that arguments diverge sharply. For
the theory of abandonment option, the relevance of this variable is inversely related
to the financial “health” of the company. Its relevance is sustained as a proxy to the
liquidation value. The abandonment option exercise by shareholders tends to occur
in a context where the company records successive losses, and investors tend to
immediately incorporate the highest probability of bankruptcy that those firms incur
(Hayn 1995; Berger et al. 1996; Subramanyam and Wild 1996; Collins et al. 1997;
Burgsthaler and Dichev 1997; Barth et al. 1998; Collins et al. 1999).
In turn, given the most recent investigations post 90s that document a larger
number of small businesses operating in high-tech sectors which report losses of
greater magnitude and for longer periods, the variable “net income—losses” has
weak predictive power of future profitability. Thus, the principle of continuity
suggests that losses are not sustainable indefinitely. According to the OM, the BVE
variable measures the stock assets held by the company which is a proxy for future
normal results. Given the definition of permanent results by Ohlson (1995), in the
medium term it is expected that the company will generate a rate of return equal to
the cost of capital. The losses are well supported in the conservatism accounting
effect, given the investment profile that characterizes this type of company—in-
vestments in intangibles.
Hence, given the principle of continuity (note that the number of bankruptcies in
the samples under study was lower than 5%), and assuming that the persistence of
losses incurred by the companies of the samples under study is a reflection of the
16
We claim again that, given the principle of continuity, it is not expected that the effect of the
extraordinary items persists in the medium and long term.
effect of accounting conservatism (based on the values assigned to the variables

R&D and R&D/sales), the variable results is somewhat related to future prof-
itability. Based on the OM, we identify the variable BVE as proxy for the normal
expected future results in the medium and long term. Thus, we formulate the second
hypothesis that held for both samples:
Hypothesis 2: The equity (BVE) is positively related to MVE
The model to estimate is the Eq. (6.6):
MVEit ¼ a0 þ a1 ðBVEÞi;t1 þ a2 ðRes IExtÞit þ eit ð6:6Þ
with i = 1, 2 … N and t = 1996–2003.

Since we include in the sample companies with negative BVE, we do not
anticipate a priori the impact of these observations on the coefficients to be esti-
mated (i.e. the variables results or BVE). As modelled by Feltham and Ohlson
(1995), the financial statements do not recognize the net present value
(NPV) associated with intangible assets, thus, the unrecorded goodwill which is a
results of the differential between the market value and book value of equity, tends
to be very high (and persistent) in technology-based companies, as a consequence
of a dual effect:
(i) the understatement of assets as a result of the investment in intangible assets
being treatment as costs and;
(ii) the fact that the market, particularly given the emerging nature of the Internet
sector, assumes a higher probability of the existence of future growth options
and, thus, has higher expectations for abnormal returns.17
Thus, the implementation of these investments can be interpreted as the exercise
of an option (call option) by the company in order to undertake growth opportu-
nities held in their portfolio.18 Assuming the principle of rationality, the exercise
(following) of the purchase options are a clear signal to the market that managers
only exercise the options if the options are in-the-money.
17
We recall that Eq. 2.23 analytically defines unrecorded goodwill as:
Pt bvt ¼ gt ¼ a1 oxat þ a2 oat þ b vt . Thus, the underestimation effect is captured by the
parameter a2 (oat—operating assets). The impact of the existence of a greater or lesser number
of growth opportunities in the portfolio is reflected in the parameter a1, which measures
expectations about abnormal profitability, and b, which reflects all information other than
financial that is coming to market, and which is useful to investors in formulating their
expectations about the growth of this sector (companies), which is immediately reflected in stock
prices, but only subsequently reported in the financial statements.
18
We emphasize once again, and with reference to the investment model of general equilibrium
(e.g., Hugonnier et al. 2005), that the option of deferral of certain investments (e.g., investment in
information technologies) can “destroy” the value of these investments, even considering a sce-
nario of moderate risk aversion.
130 6 Method
Based on this framework, and given the empirical results of Chauvin and
Hirschey (1993), Connolly and Hirschey (1984) and Hirschey (1982) for the 80s,
Joos and Plesko (2005) have analysed one sample of net firms throughout the 90s
and show that the market seems to positively value the investment in this type of
item in clear anticipation of the net present value (NPV) by disaggregating the
variable “results” into its constituents. With this breakdown we aim first to examine
whether or not the effect “positive valuation of the losses” is the result of con-
servatism accounting, and thus the adequacy of the variables R&D and advertising
as proxies for growth opportunities for these companies. Thus, given the stage of
the life cycle of the companies, whose proxy is the variable “results”, and in line
with the investment opportunity hypothesis (Szewczyk et al. 1996) and the
assessment by the market, of the asymmetrical relationship of the items R&D and
advertising, the company makes a profit or loss; so for profitable companies,
already in a stage of maturity or steady growth, the market interprets that the
“profits” reflects the present value (PV) of past investments, consequently the
indirect effect associated with this type of investment dominates (Sougiannis 1994).
For companies reporting losses in the start-up phase, losses are associated with
(successive) investments in R&D and advertising (conservatism accounting).
Investment in these items aims to increase the potential provided by the network-
effect space [World Wide Web—(www)], which is creating new growth opportu-
nities on a global scale, in order to obtain increasing returns, as modelled by Noe
and Parker (2005).
In this context, we formulated the third hypothesis:
Hypothesis 3a: There is a positive relationship between the variables R&D and
advertising and the market value of firms in the start-up phase
when firms report losses (net and non-net firms)
Hypothesis 3b: There is a negative relationship between the variables R&D and
advertising and the market value of firms in stable
growth/maturity (net and non-net firms)
To test these hypotheses, the study empirically adjusted the variable results
before extraordinary items to the values recorded by the variables R&D and
advertising. Hence the Eq. (6.6) is defined:
MVEit ¼ a0 þ a1 ðBVEÞi;t1 þ a2 ðRes I&DÞit þ a3 ðR&DÞit þ eit and ð6:7Þ
MVEit ¼ a0 þ a1 ðBVEÞi;t1 þ a2 ðRes AdvÞit þ a3 ðAdvÞit þ eit ð6:8Þ
where the variables “Res_I&D” and “Res_Adv” measure the value of the variable
results prior to the investment in R&D and advertising, respectively, and the index i
identifies the number of firms in the sample under analysis (i = 1, 2 … N) and t the
period under study (1996–2003).
Since our aim is to analyse the relationship between market capitalization and
losses incurred by the companies, and assuming that these are a result of large
investments in intangible assets (conservatism accounting), our choice was to adjust
the results to the amounts invested by year in those variables (i.e. R&D and
advertising).
In this context, we exclude the possibility of including in the analysis lagged
values for these variables because: (i) companies in the maturity/stable growth NPV
of the investments made in the past are already reflected in the variable results (the
indirect effect dominates—Sougiannis 1994); (ii) in younger companies that reg-
ister losses, the current investment in these variables acts as a proxy for growth
opportunities held in their portfolio; (iii) given this type of investment is considered
in full expenses in the period they occur, they are not subject to any impairment
tests (comparison between the cost of acquisition/production value and the mar-
ket value); and (iv) any additional information is not provided to investors in the
following years. Therefore, we argue that it is based on the current values of these
variables (i.e. R&D and advertising) that investors formulate their expectations
about the magnitude of future cash flows and the level of risk that are associated
with them.
Furthermore, given the principle of continuity (going concern), losses are not
sustainable indefinitely. At any given moment in time, it is expected that companies
manage to have earnings. Based on this reasoning, and given the period under
analysis (NEP) which now involves 8 years, the fourth hypothesis states:
Hypothesis 4: The market evaluates differently the determinants of firm value as
the sector/companies moving towards maturity
We emphasize the appropriateness of the methodology of Fama and MacBeth
(1973) to test these hypotheses, since it allows the null hypothesis that the mean of
the time series of the estimated coefficients over time is zero.
Finally, and in order to contrast the results between the two samples analysed,
net firms and non-net firms, the fifth hypothesis is:
Hypothesis 5: The variations in the market value of the net and non-net equity
firms are explained by the same determinants: “results”, “BVE”,
“R&D” and “advertising”
This hypothesis is particularly relevant because it allows us to test the extent of
the fashion (fad) effect documented by Lee (2001), Cooper et al. (2001) and Ofek
and Richardson (2003) which shows a greater expression of the phenomenon
positive valuation of the losses in the Internet sector.
In the next chapter, we conduct the empirical analysis and discuss the results.
132 6 Method
Appendix 6.1: The CUSUM Test for the Variable “Results

Res_IExt”—Net Firms
Appendix 6.2: The CUSUM Test for the Variable … 133
Appendix 6.2: The CUSUM Test for the Variable

“MVE/BVE”—Net Firms
134 6 Method
Appendix 6.3: The CUSUM Test for the Variable “Results

Res_IExt”—Non-net Firms
Results (Res_IExt) Mean–non-net firms
Results (Res_IExt) Median –non-net firms

1.6
1.2
0.8
0.4
0.0
-0.4
1998 1999 2000 2001 2002 2003
CUSUM of Squares 5% Significance

Appendix 6.4: The CUSUM Test for the Variable … 135
Appendix 6.4: The CUSUM Test for the Variable

“MVE/BVE”—Net Firms
“MVE/BVE” Mean–non net firms
“MVE/BVE” Median–Non Net firms

136 6 Method
Appendix 6.5: Percentage of Firms with R&D Expenses

Greater than Sales
Sample 1996 1997 1998 1999 2000 2001 2002 2003

Non-net firms: 62 82 92 174 250 262 229 190
No. Firms: 9 14 14 25 54 53 47 34
R&D/SALES > 1
% Firms: 14.52% 17.07% 15.22% 33.78% 21.60% 20.23% 20.52% 17.89%
R&D/SALES > 1
Net firms: 37 66 107 350 453 392 313 206
No. Firms: 6 10 10 26 32 27 20 13
R&D/SALES > 1
% Firms: 16.22% 15.15% 9.35% 7.43% 7.06% 6.89% 6.39% 6.31%
R&D/SALES > 1
References
income as a function of financial health. J Account Econ 25:1–34
Berger PG, Ofek E, Swary I (1996) Investor valuation of the abandonment option. J Financ Econ
42:257–287
Brealey RA, Myers S (2003) Princípios de Finanças Empresariais, 7th edn. McGraw-Hill, Portugal
Burgstahler DC, Dichev I (1997) Earnings, adaptation and equity value. Account Rev 72(2):187–
215
Chambers DJ (1997) The information content of negative earnings and its relation with initial-loss
persistence. Working Paper, University of Illinois at Urbana-Champaign. http://ssrn.com/
abstract=15048
Chauvin KW, Hirschey M (1993) Advertising, R&D and the market value of the firm. Financ
Manag 22(4):128–140
values over the past forty years. J Account Econ 24(1):39–67
of equity. Account Rev 74(1):29–61
Connolly RA, Hirschey M (1984) R&D, market structure, and profits: a value-based approach.
Rev Econ Stat 66(4):682–687
Cooper MJ, Dimitrov O, Rau R (2001) A rose.com by any other name. J Finance 56(6):2371–2388
Damodaran A (2001) The dark side of valuation. Prentice Hall, USA
Demers E, Lev B (2001) A rude awakening: the internet shakeout in 2000. Rev Account Stud 60(2
& 3):331–359
Easton PD, Sommers GA (2003) Scale and the scale effect in market-based accounting research.
J Bus Finance Account 30(1&2):25–56
Eviews Manual (2004) Eviews 5 user’s guide. Quantitative Micro Software, UK
Fama E, French K (1998) Taxes, financing decisions, and firm value. J Finance 53(3):819–842
References 137
Fama E, MacBeth J (1973) Risk, return, and equilibrium: empirical tests. J Political Econ 81:607–
636
activities. Contemp Account Res 11(2):689–731
Francis J, Schipper K (1999) Have financial statements loss their relevance? J Account Res 37
(2):319–352
Greene WH (2008) Econometric analysis, 6th edn. Pearson Prentice Hall, Upper Saddle River
Gujarati DN (2002) Basic econometrics, 4th edn. McGraw-Hill, New York
Hayn C (1995) The information content of losses. J Account Econ 20(2):125–153
Hirschey M (1982) Intangible capital aspects of advertising and R&D expenditures. J Ind Econ 34
(4):375–390
Hugonnier J, Morellec E, Sundaresan SM (2005) Irreversible investment in general equilibrium.
Working Paper no. 230. Columbia University: National Centre of Competence in Research
Financial Valuation and Risk Management
Johnston J, Dinardo J (2001) Métodos Econométricos, 4th edn. McGraw-Hill, New York
Joos P, Plesko GA (2005) Valuing loss firms. Account Rev 80(3):847–870
Keating EK, Lys TZ, Magee RP (2003) Internet downturn: finding valuation factors in Spring
2000. J Account Econ 34(1&3):189–326
Lee PM (2001) What’s in a name.com? The effects of “.com” name changes on stock prices and
Lev B, Sougiannis T (1996) The capitalization, amortization, and value-relevance of R&D.
J Account Econ 21(1):107–138
firms. Paper presented at 14th annual conference on Financial Economics and Accounting
(4):411–433
Noe T, Parker G (2005) Winner take all: competition, strategy, and the structure of returns in the
internet economy. J Econ Manag Strategy 14(1):141
Ofek E, Richardson M (2003) Dot.Com Mania: the rise and fall of internet stocks prices. J Finance
58(3):1113–1138
Ohlson J (1995) Earnings, book values, and dividends in equity valuation. Contemp Account Rev
11(2):661–687
Penman SH (2003) Financial statement analysis and security valuation, 2nd edn. McGraw-Hill, US
Schultz P, Zaman M (2001) Do the individuals close to internet firms believe they are overvalued.
J Financ Econ 59(3):347–381
Sougiannis T (1994) The accounting based valuation of corporate R&D. Accounting Review 69
(1):44–68
Subramanyam KR, Wild JJ (1996) Going-Concern status, earnings persistence, and informative-
ness of earnings. Contemporary Accounting Research 13(1):251–273
Szewczyk SH, Tsetsekos GP, Zantout Z (1996) The valuation of corporate R&D expenditures:
evidence from investment opportunities and free cash flow. Financ Manag 25(1):105–110
Commer 4(4):23–40. Doi:10.1300/J179v04n04_02
stocks. J Account Res 38(Supplement):137–162
Chapter 7
Analysis and Discussion of Results
Abstract Given the magnitude and persistence of losses reported by the companies
under study, this chapter aims to test empirically the phenomenon of “positive
valuation of the losses” based on the theoretical framework of the OM and FOM
models. The objective is to understand how investors assess this type of company,
given the persistence and the continuous reporting of losses. So, after some con-
siderations on the econometric procedures, we analyze the results obtained with
reference to each of the research hypotheses formulated.
Keywords Persistence of losses

Positive valuation of losses Ohlson and

Feltham and Ohlson models Empirical results
7.1 Discussion and Analysis of Results
7.1.1 Econometric Aspects
We begin by estimating the Eq. 6.5 for both samples and for the two subgroups of
companies: companies with profits and companies with losses. Table 7.1 presents
the results of regressions year by year for the net firms, with the aim of estimating
the aggregate coefficients according to the methodology of Fama and MacBeth
(1973). Because the estimated coefficients could be correlated over time, we used
the Ljung-Box statistic tests.1 If the estimated coefficients show a positive auto-
P 2
1
The Box-Ljung statistic is defined as: LB ¼ nðn þ 2Þ m k¼1 q^k =n k v2m , where m is the
number of degrees of freedom, k is the number of autocorrelation coefficients to calculate, and
n the number of observations (Ljung and Box 1978).

DOI 10.1007/978-981-10-3009-3_7
140 7 Analysis and Discussion of Results
correlation, and as suggested by Core et al. (2003), the t-statistic was calculated by
adjusting the standard deviation by the method of Newey and West (1987)
assuming six lags.2
With respect to the annual regressions estimations: (i) whenever we detected the
heteroscedasticity, this effect was corrected using the White (1980) test; (ii) with
respect to multicollinearity, we apply the variance inflation factor (VIF) and the
values are less than ten; (iii) the Durbin–Watson statistic (DW) shows the absence
of autocorrelation; and (iv) with regards to the assumption of normality, whenever
the Jarque–Bera test (JB) was rejected, the Kolmogorov–Smirnov (KS) is applied.3
Next we present and discuss the results:
7.1.2 Validation of Empirical Research Hypotheses
7.1.2.1 Validation of Research Hypothesis 1
Regarding the first hypothesis the results are reported in Table 7.1.
The results reported in Table 7.1 show that the coefficient of the variable results
(Res_IExt: earnings before extraordinary items—annual item Compustat 237) in the
sample of net firms with losses is negative and statistically significant at a 5% level.
Because this subsample registers consecutive losses, the market seems to positively
value the losses incurred. Notice that the number of companies with losses over
time is persistent (and the most significant subgroup). Thus, the first hypothesis is
accepted.
When comparing the results obtained with the sample of non-net firms
(Table 7.2), the variable earnings before extraordinary items (annual item
Compustat 237) report a negative value as expected, however with a lower statis-
tical significance. The empirical validation of the phenomenon “positive valuation
of losses” is not consensual, even in the universe of net firms. For example,
Trueman et al. (2000), Martinez and Clement (2002), Hand (2003) and Tockic
(2005) empirically support this phenomenon throughout the period studied [these
authors cover all the period of the crash except Trueman et al. (2000)]. Hand
(2001), using a sample split according to the variable web traffic, only documents
this phenomenon in the pre-crash period like Kozberg (2009) and Demers and Lev
(2001). In the post-crash the variable reverses in sign. However, out of the universe
of net firms, very few studies on this topic report this phenomenon. One example is
2
Similar to the White (1980) test, which does not need prior knowledge about the pattern of
heteroscedasticity, the Newey and West (1987) method is robust in the presence of autocorrelation.
If the residuals are not autocorrelated, the adjustment of standard deviations by the Newey and
West (1987) method is similar to the method of ordinary least squares. According to Greene
(2000), we estimate the standard deviations assuming six lags on the assumption that residuals are
not correlated over six periods.
3
For detail information about the tests: DW, JB and KS, see Gujarati (2002).
7.1 Discussion and Analysis of Results 141
Table 7.1 Annual regressions for the sample of net firms (model 6.5)
Sample with profits Sample with losses
#Na ab0 Res_IExtb R2 Adj. (%) # Na ab0 Res_IExtb R2 Adj. (%)
1996 34 5.05 ***
0.05***
44.26 37 4.08 ***
−0.026 ***
16.23
(2) (28:20) (5.215) (0) (17.195) (−3.838)
1997 49 5.06*** 0.037*** 40.31 66 3.94*** −0.022*** 19.85
(1) (30.462) (5.781) (0) (16.762) (−2.978)
1998 47 4.84*** 0.047*** 27.24 107 4.67*** −0.014*** 14.47
(1) (21.714) (4.268) (4) (25.452) (−4.351)
1999 80 6.44*** 0.02*** 12.68 350 6.19*** −0.005 6.94
(5) (31.475) (3.532) (16) (59.44) (−4.227)
2000 74 5.08*** 0.022*** 20.83 453 4.42*** −0.002*** 9.81
(2) (25.246) (4.859) (3) (45.728) (−4.065)
2001 36 5.40*** 0.027*** 32.93 392 4.15*** −0.001*** 8.20
(1) (21.72) (4.264) (10) (39.645) (−5.484)
2002 54 4.61*** 0.03*** 35.29 313 3.69** −0.002*** 8.42
(5) (18.276) (5.468) (12) (34.438) (−3.318)
2003 101 5.50*** 0.017*** 36.50 206 4.07*** −0.007*** 12.20
(5) (37.112) (7.647) (15) (30.818) (−5.43)
Meanc 5.25*** 0.031*** 31.25 4.402*** −0.01** 12.02
T-Statistics d
26.545 7.087 15.93 −2859
Estimated model
MVEit = a0 + a1 (Res_IExt)it + eit
where MVE represents the market value of the company and Res_Ext the results before extraordinary items
a
Number of observations with negative BVE for the variable value in brackets
b
T-statistic for the estimated coefficients from year to year in brackets
c
Corresponds to the estimated coefficients average, including the constant
d
The t-statistic (two-tailed test), calculated from the ratio of the average standard deviation multiplied by 8½
(***) and (**) indicate significance of 1 and 5% level, respectively
the study by Collins et al. (1997) for the period 1953–1993; Collins et al. (1999)
and Frazen and Radhakrishnan (2009) obtained a R2 of 9 and 6% respectively, very
similar to values reported by us for the sample of net firms (12.02%) and non-net
firms (10.30%). Core et al. (2003) also noted this phenomenon in the two samples:
high-technology companies and start-up companies.
These results and the cluster effect that characterized the IPOs in both samples
support the first hypothesis, i.e. younger firms, particularly technology-based ones,
with negative results are positively valued by the market. Thus, the market appears
to associate losses with the high investments in intangible assets, hence the effect of
conservatism accounting is compensated for by an overestimation of expected
abnormal results.
The lower statistical significance for the sample of non-net firms can be
explained by the fact that in spite of the number of companies with losses persisting
over time, the difference between the number of firms reporting profits and losses is
not so marked as in the sample of net firms, as can be seen in Fig. 7.1. Moreover,
when comparing the magnitude of losses reported by the two samples, the values
Table 7.2 Annual regressions for the sample of non-net firms (model 6.5)
Year Sample with profits Sample with losses
# ab0 Res_IExtb R2 Adj. # ab0 Res_IExt2 R2 Adj.
Na (%) Na (%)
1996 81 4.23*** 0.07** 40.20 62 3.46*** −0.078*** 26.90
(2) (32.02) (7.065) (3) (17.761) (−4.844)
1997 109 4.62*** 0.033*** 26.37 82 3.72*** −0015 5.30
(5) (37.216) (3.107) (2) (17.725) (−1.647)
1998 120 4.43*** 0.034*** 42.85 92 3.84*** −0.005 1.00
(0) (42.447) (9.499) (1) (22.658) (−1301)
1999 182 4.52*** 0.051*** 25.34 174 4.74*** −0.012** 9.00
(5) (35.439) (5.772) (9) (25.706) (−2124)
2000 195 4.18*** 0.058*** 38.14 250 4.52*** −0.003** 4.20
(6) (31.46) (8.699) (6) (32.028) (−2.029)
2001 140 4.52*** 0.060*** 47.79 262 4.28*** −0.004*** 11.74
(4) (34.919) (10.152) (10) (37.117) (−5975)
2002 145 4.44*** 0.036*** 40.12 229 3.83*** −0.002** 5.70
(3) (34.68) (5.691) (10) (29.525) (−2.201)
2003 146 5.09*** 0.023*** 36.52 190 4.34*** −0.011*** 18.54
(4) (39.077) (5.700) (12) (31.365) (−5.122)
Meanc 4.505*** 0.0456***
37.17 10.30 4.09*** −0.016
T-Statistics 45.437 7.87 25.995 −1.832
td
Estimated model
MVEit = a0 + a1 (Res_IExt)it + eit
where MVE represents the market value of the company and Res_Ext the results before
extraordinary * items
a
b
c
Corresponds to the estimated coefficients average, including the constant
d
The t-statistic (two-tailed test), calculated from the ratio of the average standard deviation
multiplied by 8½
(***) and (**) indicate significance 1 and 5% level, respectively
Sample of net firms Sample of non-net firms

600 500
450
500
400
400 350
300
300 250
200
200
150
100
100
50
0 0
1996 1997 1998 1999 2000 2001 2002 2003 1996 1997 1998 1999 2000 2001 2002 2003
Total firms Firms with porfits Total firms Firms with porfits
Firms with losses Firms with losses
Fig. 7.1 a, b Evolution of the number of firms per sample depending on the reported value of the
variable results
are substantially lower. In the sample of net firms the values of the mean and
median value for this variable are, respectively, 96.9 and 24.29 million dollars
(panel data) (Table 6.3). In the sample of non-net firms (Table 6.4), the losses
amounted to 44.18 and 14.67 million dollars for the mean and median, respectively
(panel data). These figures allow us to associate a fashionable effect (fad) (e.g.,
Bartov et al. 2002) that is more pronounced in the sample of net firms, which Jorion
and Talmor (2006) justified due the high growth opportunities associated with net
firms, given the emerging nature of this sector.
We recall that when we use the variable results before extraordinary items, we
are violating the CSR assumption of the OM, which relies on the overall net income
of the company—comprehensive income.4 However, similarly to Dechow et al.
(1999) and Collins et al. (1999), the effect of extraordinary items is a short-term
one. Indeed, if we estimate the model 6.5 using the net income variable (annual item
Compustat 172), the results do not change.5
Regarding the group with profits in both samples, and as expected, the
explanatory power of the model is higher (i.e. adjusted R2) compared to the group
with losses (31.25 vs. 12.02% in the sample of net firms and 37.17% against
10.30% in the sample of non-net firms). These results confirm that the explanatory
power of positive results is enhanced by their persistence (Miller and Modigliani
1961). As argued by Basu (1997), the effect of conservatism accounting, and given
the principle of prudence, generates an asymmetric effect of evaluation because
losses (bad news) tend to be recognized immediately in the period in which they
occur, thus future results are protected from bad news. As for profits, these tend to
be more persistent and lasting over time, since they are recognized only when they
are effective and measured objectively.
The phenomenon “positive valuation of the losses” is a new one, highly prevalent
in the 1990s. Some authors argue that the results obtained so far are controversial
and not robust resulting from an incorrect specification of the evaluation model.
Collins et al. (1999) argue that this phenomenon is the result of omitted variables,
4
According to the FASB (Financial Accounting Standard Board), comprehensive income is
defined as: “… The change in equity (net assets) from operations … and other events and
circumstances from non-owner sources. It includes all changes in equity except during the period
those resulting from investments by owners and distributions to owners” (para 70).
5
For the sample of net firms, R2 stands at 11.89%, and the estimated coefficient for the variable
results (−0.0099) is statistically significant at 5% level in the group with losses. In the group with
profits, the R2 amounts to 27.68% and the estimated coefficient (0.028), revealing significance at
1% level. Similar results are recorded in the sample of non-net firms: R2 reported a value of 9.48%
in the group with losses, and the estimated coefficient (−0.015) is not statistically significant. In the
group with profits, R2 achieves a mean of 33.95%, with an estimated coefficient (0.0428) sig-
nificant at a 1% level.
Table 7.3 Correlation matrix—sample of net firms

Panel A: Companies with losses (n = 1.924)
MVE BVE RL Res_IExt R&D Adv
MVE 1 0.588** −0.300** −0.311** 0.327** 0.111**
BVE 0.404**
1 −0.717 **
−0.699 **
0.385**
0.190**
NI −0.150 **
−0.666 **
1 0.976**
0.323**
−0.232**
Res_IExt −0.155 **
−0.675 **
0.990 **
1 −0.320 **
−0.236**
R&D 0.262**
0.381**
−0.379 **
−0.389 **
1 0.132**
Adv 0.133**
0.152**
−0.19 **
−0.192 **
0.084**
1
Panel B: Companies with profits (n = 475)
MVE 1 0.676** 0.627** 0.650** 0.407** 0.241**
BVE 0.482** 1 0.539** 0.585** 0.377** 0.193**
** ** ** **
NI 0.380 0.622 1 0.967 0.216 0.174**
** ** ** **
Res_IExt 0.395 0.715 0.936 1 0.217 0.181**
** ** ** **
R&D 0.456 0.537 0.466 0.490 1 0.282**
** ** ** ** **
Adv 0.507 0.737 0.590 0.642 0.526 1
**
Correlation is significant at the 5% level. Numbers below the diagonal represent Pearson
correlations. Spearman’s correlations appear above the diagonal [For a definition of the coefficients
of variation of Pearson and Spearman’s correlations, see Gujarati (2002)]. The correlation matrix
was calculated according to a panel data structure. MVE corresponds to the market value of equity,
BVE is the book value of equity, NI and RL Res_IExt corresponds to the amount of net income
and earnings before extraordinary items. R&D and Adv report the amount invested in research and
development and advertising, respectively
explicitly referring to the omission of the BVE variable from the model 6.5. The
argument is sustained by the fact that the BVE variable is positively correlated with
MVE (dependent variable) and negatively (positive) with the results (independent
variable). Therefore omitting this variable from the evaluation model induces a
negative (positive) bias to the estimated coefficients for the variable results.6
Thus, in order to validate empirically the arguments of Collins et al. (1999), we
examine the correlation matrix for each sample and its subgroups. The results are
reported in Tables 7.3 and 7.4 and confirm the results predicted by Collins et al.
(1999).
For the group with profits (net and non-net firms) the correlation between the
BVE and all the other variables is positive and statistically significant at a 5% level.
For the subsample with losses, the correlation of BVE with the variable results
(Res_IExt) changes the sign, i.e. becomes negative, keeping a statistical
6
Considering the OM (model 6.3), omitting for simplicity theh index referring to ithe company
(i) MVEt ¼ a þ a1 ðRes IExtÞt þ a2 ðBVEÞt1 þ et Eða1 Þ ¼ a1
covðRes IExt;BVEÞ a . Then,
varðRes IExtÞ 2
the coefficient a1 depends on the covariance between the variable “results—Res_IExt” and the
variable “BVE” and the coefficient a2, which measures the relationship between the dependent
variable (MVE) and the independent variable BVE (Greene 2000).
Table 7.4 Correlation matrix—sample of non-net firms

Panel A: Companies with losses (n = 1.341)
MVE 1 0.714** −0.445** −0.473** 0.443** −0.09**
BVE 0.495**
1 −0.665 **
−0.651 **
0.381 **
−0.065**
NI −0.183 **
−0.633 **
1 0.962 **
−0.347 **
−0.015
Res_IExt −0.198** −0.628** 0.965** 1 −0.350** −0.021
R&D 0.188** 0.390** −0.184** −0.183** 1 −0.0133**
Adv 0.098** 0.124** −0.384** −0.385** −0.039 1
Panel B: Companies with profits (n = 1.118)
MVE BVE RL Res_Ext R&D Adv
MVE 1 0.732** 0.693** 0.700** 0.345** 0.031
** ** ** **
BVE 0.515 1 0.609 0.632 0.193 0.051
NI 0.626** 0.508** 1 0.972** 0.102** 0.073**
** ** ** **
Res_IExt 0.678 0.599 0.892 1 0.095 0.072**
** ** ** **
R&D 0.629 0.519 0.553 0.599 1 0.011
Adv 0.003 0.019 0.022 0.023 −0.007 1
**
Correlation is significant at 5% level. Numbers below the diagonal represent Pearson
correlations. Spearman’s correlations appear above the diagonal [For a definition of the
coefficients of variation of Pearson and Spearman’s correlations, see Gujarati (2002)]. The
correlation matrix was calculated according a panel data structure. MVE corresponds to the market
value of equity, BVE is the book value of equity, NI and Res_IExt corresponds to the amount of
net income and earnings before Extraordinary items, R&D and Adv report the amount invested in
research and development and advertising, respectively
significance of a 5% level. Given these results, the next step was to introduce the
variable BVE into the valuation model (model 6.3).
Tables 7.5 and 7.6 show the results for both samples and both groups.
Focusing the analysis on the group with losses, with the introduction of the
variable BVE in the evaluation model, the effect of positive valuation of losses
disappears in the sample of net firms. These results are partially consistent with the
results of Collins et al. (1999), for whom the phenomenon of valuation of losses is
the result of an incorrect specification of the evaluation model, referring specifically
to the capitalization results model. This conclusion is even stronger when we
analyze the results for the sample of non-net firms. Figure 7.2 shows how the
omission of the variable BVE induces a negative (positive) bias on the estimated
coefficient for the variable results when it records losses (profits).
The inclusion of the variable BVE increases the explanatory power for both
groups. For the sample of net firms the increase in the adjusted R2 is 14.63% (the
increase is 12.02–26.65%) compared to 21.59% in sample of non-net firms (R2 goes
from 10.30 to 31.89%), which indicates that this variable has an incremental
explanatory power for firms with losses, in addition to the variable results.
Thus, given the increase in the explanatory power of the evaluation model after
inclusion of the variable BVE in the group with losses, and that: (i) in both samples
146

# Na ab0 BVEb Res_IExtb Adj. R2 (%) # Na ab0 BVEb Res_IExtb Adj. R2 (%)
1996 34 5.03*** 0.0005 0.049*** 42.79 37 3.45*** 0.022*** −0.007 46.53
(2) (27.07) (0.426) (4.682) (0) (13.589) (3.664) (−0.471)
1997 49 4.89*** 0.003** 0.032*** 45.92 66 3.79*** 0.006* −0.012*** 28.63
(1) (28.278) (2.424) (4.96) (0) (16.799) (1.856) (−3.194)
1998 47 4.59*** 0.008*** 0.012 37.51 107 4.33*** 0.006*** −0.008** 30.88
(1) (20.551) (2.898) (0.781) (4) (24.347) (5.093) (−2.533)
1999 80 6.17*** 0.0028** 0.001** 28.74 350 6.12*** 0.001*** −0.004*** 12.03
(5) (28.276) (2.412) (2.062) (16) (62.868) (4.597) (−3.813)
2000 74 4.79*** 0.002*** 0.013*** 35.77 453 4.38*** 0.001*** −0.0002 16.80
(2) (19.526) (4.612) (2.928) (3) (46.144) (3.072) (−0.59)
2001 36 5.23*** 0.002*** 0.016*** 42.15 392 3.92*** 0.002*** 0.001** 23.82
(1) (21.683) (2.534) (2.237) (10) (35.277) (3.937) (1.98)
2002 54 4.64*** 0.001** 0.016* 36.51 313 3.47*** 0.003*** 0.003** 24.91
(5) (16.392) (2.221) (1.959) (12) (27.47) (3.549) (2.423)
2003 101 5.50*** 0.0004 0.014** 36.74 206 3.94*** 0.004*** 0.002 29.54
(5) (34.726) (0.639) (2.195) (15) (30.24) (7.214) (1.176)
Meanc 5.11*** 0.0024** 0.02*** 38.27 4.174*** 0.005** −0.003 26.65
T-Statisticsd 27.364 2.897 4.271 13.795 2.21 −1.598
Estimated model
MVEit = a0 + a1 (BVE)it + a2 (Res_IExt)it + eit
where MVE represents the market value of the company, BVE the book value of equity and Res_IExt the results before extraordinary items
a
b
c
Corresponds to the average of the estimated coefficients, including the constant
d
(***), (**) and (*) indicate significance level of 1, 5 and 10%, respectively
7 Analysis and Discussion of Results
# Na ab0 BVEb Res_IExtb Adj. R2 (%) # Na a0b BVEb Res_IExtb Adj. R2 (%)
1996 81 3.94*** 0.011*** 0.0354*** 55.83 62 3.23*** 0.014*** −0.008 49.82
(2) (34.521) (5.381) (3.412) (3) (19.331) (5.328) (−0.422)
1997 109 4.19*** 0.011*** 0.003 47.80 82 3.30*** 0.016*** 0.012 35
(5) (37.474) (6.706) (0.455) (2) (13.147) (3.271) (1.45)
1998 120 4.17*** 0.006*** 0.018*** 50.28 92 2.87*** 0.023*** 0.014*** 44.15
(0) (34.661) (3.058) (3.058) (1) (16.749) (8.422) (3.586)
1999 182 4.39*** 0.003*** 0.04*** 29.06 174 4.51*** 0.003*** −0.008** 22.18
(5) (32.935) (3.179) (4.185) (9) (25.981) (2.592) (−2.212)
2000 195 3.95*** 0.005*** 0.039*** 44.07 250 4.22*** 0.002*** −0.0001 20.46
(6) (28.511) (3.615) (4.794) (6) (28.183) (5.09) (−0.094)

2001 140 4.35*** 0.003*** 0.046*** 54.20 262 4.08*** 0.003*** 0.020** 28.54
(4) (36.488) (4.199) (11.236) (10) (34.220) (6.199) (2.464)
2002 145 4.25*** 0.003*** 0.025*** 48.89 229 3.52*** 0.003*** 0.002*** 25.74
(3) (33.963) (3.339) (4.034) (10) (29.332) (6.062) (2.85)
2003 146 4.94*** 0.002*** 0.016*** 44.21 190 4.12*** 0.002*** −0.005*** 29.22
(4) (36.664) (4.827) (3.532) (12) (31.341) (5.059) (−2.65)
Meanc 4.274*** 0.005*** 0.028*** 46.79 3.74*** 0.008** 0.003 31.89
T-statisticsd 38.181 4.117 5.235 17.993 2.841 0.867
Estimated model
MVEit = a0 + a1 (BVE)it + a2 (Res_IExt)it + eit
where MVE represents the market value of the company, BVE the book value of equity and Res_IExt the results before extraordinary items
a
b
c
d
(***), (**) and (*) indicate significance of a 1, 5 and 10% level, respectively
147
Sample of net firms Sample of non-net firms

30 35
25 30
25
20
20
15
15
10
10
5 5
0 0
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Modelo 6.5 Modelo 6,3 Modelo 6.5 Modelo 6,3
Fig. 7.2 Relationship between the estimated coefficients after controlling for BVE variable
(model 6.3) and without the control for BVE variable (model 6.5)
the group reporting losses has more companies (Tables 5.3 and 5.4); (ii) the number
of bankruptcies recorded in the samples under study is lower (Tables 5.3 and 5.4);
and (iii) the investment profile of this group of companies is characterized by
investment in intangible assets compared to the group with profits (Tables 6.3 and
6.4),we argue that the variable BVE and, in light of the OM, is a proxy of future
normal results for the companies reporting losses in those companies—net and non-
net firms.
Moreover, given the profile of the samples under study, high-technology com-
panies (Table 5.6), i.e. those where intangible assets are considered as immediate
costs, consequently, have no value when considered alone, especially in a situation
of poor financial performance (e.g. when the company report losses) and due the
absence of organized markets for their assets (except for trademarks and patents); in
this context, the variable BVE is also a proxy for recognized assets. Thus, this
proves to be a valuable variable for reducing the costs of monitoring/control by
creditors (Jensen and Meckling 1976). With these results and according to the
theory of the abandonment option, which associates a higher probability of bank-
ruptcy with persistent losses, the market is recognizing the variable BVE as a good
proxy to the liquidation value of the firms.
With reference to the profit group (in both samples), the results are as expected.
The profits continue to be the main determinant of value for these companies, as
supported by Miller and Modigliani (1961). The additional explanatory power after
including the variable the BVE in model 6.5 is substantially lower compared to the
group with losses (in the sample of net firms the increase was 7.02%, compared to
9.62% in the sample of non-net firms). As for the variable BVE, this also has a
positive and statistically significant coefficient. Consistently positive results, and
according to the CSR principle, reinforce the value of the stock of assets held by the
company, increasing its future ability to generate positive results. Based on the
above results, we support the second hypothesis: the BVE is positively related with
MVE (for both net and non-net firms), and according to the OM is a proxy for the
company’s future results.
Given the results for the first hypothesis, the market seems to positively value the
losses reported either by net firms or non-net firms by linking them to the impact of
investments in intangible assets (conservatism accounting). For this type of com-
pany, the difference between the market value and book value of equity (un-
recorded goodwill—Tables 6.3 and 6.4) is very high. In this context, contrary to the
theory of the abandonment option, the losses are not a proxy for the highest
probability of bankruptcy which those companies incur. As modelled by Feltham
and Ohlson (1995), the unrecorded goodwill is due to the fact that the financial
statements, and in accordance with GAAP, do not recognize the value of the cash
flows associated with investment projects of R&D and advertising which are typical
in technology-based companies.
These accounting procedures tend to accentuate the conservatism accounting
effect, i.e. the undervaluation of the assets of these companies, and conse-
quently, the value of the results. But on the other hand, the market combines these
investments with the existence of larger growth opportunities, particularly in net
firms, given the emerging nature of this sector, generating great expectations about
future abnormal returns.
Thus, the implementation of these investment projects in R&D and advertising is
as a call option, and given the principle of rationality, it is expected that managers
undertake only those options that are in-the-money, i.e. they invest in projects that
are associated with medium-term expectations of abnormal profitability. It is
therefore important to disaggregate the “income” variable into its constituents, in
order to examine how the market assesses the R&D and advertising variables,
proxies for growth opportunities.
Thus we aim to empirically test how the conservatism accounting affects the
statistical relationship between the market value of the equity of net firms and the
related financial information. The sustainability of these variables as proxies for
growth opportunities is documented in the correlation matrix (Tables 7.3 and 7.4).
The Pearson and Spearman correlations among the variables MVE and R&D are
positive and statistically significant at a 5% level. For the advertising variable, the
results are more tenuous, but investigations are unanimous in associating this
variable with a short-term effect (e.g., Sougiannis 1994; Chan et al. 2001).
The results from the model 6.6 are reported in Tables 7.7 and 7.8. The findings
show that for both samples, and with reference to the group with losses, and as
expected, the variable R&D has a positive and statistically significant coefficient.
This result, in line with the investment opportunity hypothesis, indicates that the
market evaluates this variable as an asset (and not as a cost), thus anticipating the
net present value (NPV) associated with this type of investment. We can therefore
conclude that investors, for this particular group of companies, use this variable to
formulate their expectations about future cash flows and the level of risk associated
with them. Ben-Zion (1978) argued that the difference between the market value
and book value of the equity of a company (the unrecorded goodwill in OM and
150

# Na ab0 BVEb Res_R&Db R&Db Adj. R2 (%) # Na ab0 BVEb Res_R&Db R&Db Adj. R2 (%)
1996 34 4.92*** −0.0006 0.044*** 0006 45.95 37 3.45*** 0.022*** −0.007 0.002 44.97
(2) (25.492) (−0.489) (4.106) (0.176) (0) (15.337) (4.306) (−0.78) (0.091)
1997 49 4.88*** 0007 12:03*** −0.011 49.16 66 3.68*** 0.005 −0.011*** 0.047** 31.59
(1) (29.03) (1.268) (4.674) (−0.898) (0) (15.772) (1.508) (−3.741) (2.69)
1998 40 4.69*** 0.007** 0019 −0022 39.23 89 4.24*** 0.005*** −0.008** 0.023*** 29.53
(1) (19.822) (2.561) (0.856) (−0.633) (4) (21.303) (3.325) (−2.078) (2.213)
1999 64 5.94*** 0.0067*** −0.006 0018* 35.74 258 6.11*** 0.001*** −0.002 0.038*** 18.14
(3) (26.163) (4.128) (−0.884) (1.822) (11) (44.344) (3.763) (−1.424) (3.566)
2000 59 4.51*** 0.003*** 0007* 0018 34.96 343 4.53*** 0.0004*** −0.0002 0.007** 21.74
(1) (15.135) (3.44) (1.869) (1.045) (1) (41.304) (2.864) (−0.518) (2.047)
2001 36 5.15*** 0001 0019** −0.003 45.30 314 4.07*** 0.001*** 0.001** 0.006 25.24
(1) (21.636) (1.016) (2.603) (−0.324) (6) (30.274) (3.539) (2.558) (0.994)
2002 44 4.61*** 0002 0.026** −0.024 36.29 265 3.43*** 0.002*** 0.004*** 0.009 30.27
(3) (17.597) (1.364) (2.276) (−1.349) (8) (27.025) (4.133) (3.923) (1.551)
2003 88 5.39*** 0.0002 0.011** 0002 46.77 183 3.96*** 0.004*** 0.003 −0.003 30.68
(4) (39.851) (0.576) (2.475) (0.371) (13) (31.341) (6.636) (1.316) (−0.948)
Meanc 5.011*** 0.0032** 0019** −0.002 41.68 4.183*** 0.005** −0.002 0.016** 29.02
T-statisticsd 29.89 2.853 3.497 −0362 13.681 2.019 −1.288 2.52
Estimated model
MVEit = a0 + a1 (BVE)it + a2 (Res_R&D)it + a3 (R&D)it + eit
where MVE represents the market value of the company, BVE the book value of equity, Res_R&D the results before extraordinary items and R&D the amount
invested in research and development
a
Number of observations with negative values for the variable BVE in brackets
b
c
d
The t-statistic (two-tailed test) calculated from the ratio of the average standard deviation multiplied by 8½
(***), (**) and (*) indicates significance level of 1, 5 and 10%, respectively
# Na ab0 BVEb Res_R&Db R&Db Adj. R2 (%) # Na ab0 BVEb Res_R&Db R&Db Adj. R2 (%)
1996 81 3.93*** 0.011*** 0.039*** −0.04* 55.32 62 3.15*** 0.014*** 0.004 0.025 51.78
(2) (33.04) (5.359) (2.784) (−1.946) (3) (18.704) (5.462) (0.206) (1.212)
1997 109 4.19*** 0.011*** 0.0005 0.014 48.19 82 3.08*** 0.015*** 0.013 0.029* 39.75
(5) (37.616) (5.894) (0.078) (0.984) (2) (11.332) (3.1) (1.646) (1.866)
1998 120 4.17*** 0.006*** 0.01* 0.007 52.67 92 2.69*** 0.019*** 0.01*** 0.031** 50.53
(0) (37.27) (4.473) (1.713) (0.64) (1) (15.966) (6.834) (2.819) (2.357)
1999 157 4.42*** 0.003* 0.024** 0.0001 32.56 155 4.25*** 0.001*** −0.017*** 0.053*** 28.00
(4) (30.923) (1.907) (2.088) (0.004) (9) (25.833) (2.705) (−3.812) (5.472)
2000 173 4.00*** 0.005*** 0.031*** −0.022* 44.44 214 3.95*** 0.002*** −0.0002 0.033*** 30.96
(5) (26.887) (3.313) (3.931) (−1.853) (5) (23.626) (4.78) (−0.410) (4.753)
2001 59 4.94*** 0.001 0.034*** −0.025*** 61.02 183 4.1*** 0.002*** 0.002 0.019*** 38.88
(1) (31.769) (1.139) (4.801) (−2.334) (6) (24.858) (2.806) (0.002) (3.772)
2002 56 4.67*** 0.002*** 0.012* −0.005 59.10 164 3.38*** 0.001** 0.001 0.025*** 42.28
(1) (35.867) (3.361) (1.813) (−0.473) (5) (26.556) (1.902) (0.884) (5.960)
2003 60 5.29*** 0.001*** 0.002 0.007 45.64 139 4.35*** 0.001*** −0.0004 0.016*** 38.23
(3) (31.663) (2.501) (0.397) (0.844) (5) (37.03) (2.916) (−0.175) (4.373)
Meanc 4.452*** 0.005** 0.019** −0.008 49.87 3.619*** 0.007** 0.002 0.029*** 40.05
T-statisticsd 26.279 3.523 3.618 −1.12 16.506 2.533 0.471 7.075
Estimated model
MVEit = a0 + a1 (BVE)it + a2 (Res_R&D)it + a3 (R&D)it + eit
where MVE represents the market value of the company, BVE the book value of equity, Res_R&D the results before extraordinary items and R&D the amount
invested in research and development
a
b
c
d
(***), (**) and (*) indicate a significance level of 1, 5 and 10%, respectively
151
FOM terminology) was positively associated with the ratio of R&D over sales; this
ratio measures the intensity of investments in intangible assets.
Concerning the variable results adjusted for R&D expenses “Res_R&D”, and the
similarity to the model 6.3, the coefficient of this variable remains negative but not
statistically significant in the sample of net firms. In the group non-net firms, the
coefficient of this variable reverses the sign, i.e. it is positive and not statistically
significant. These findings suggest that conservatism accounting remains for net
firms even after we adjust the variable results for investments in R&D. The results
of the variable BVE show a positive and statistically significant coefficient, indi-
cating, as predicted by the OM, that BVE is a good proxy for the value of the assets
necessary for the company to undertake in the future the growth opportunities in its
portfolio.
The findings also show that after the breakdown of the variable “results” into its
components, the increase of the explanatory power of the model 6.6, is higher in the
subsample of non-net firms 8.16% (i.e. 31.89–40.05%, model 6.3) compared to
2.37% in the sample of net firms (26.65–29.02%, model 6.3). We can sustain this
result by the fact that the values of the variable R&D are underestimated in order to
preserve the sample size; when the values of this variable are not available (na—not
available), we assume zero. As to the group with profits in both samples, the
asymmetric evaluation of variable R&D is confirmed by the market, i.e. the coef-
ficient of the variable is negative but not statistically significant. For the group with
profits in both samples, the primary determinant of value is the profit variable
(Res_R&D), significant at a level of a 5%. As for the variable R&D, the coefficient
is negative, indicating that this group of companies is associated with a more stable
phase. The variable results already reflect the net present value of the investments
made in the past, so having a predominately indirect effect, as suggested by
Sougiannis (1994). The weak statistical significance for this variable is explained
due to the fact that these companies are very young. Tables 6.3 and 6.4 show that
the differences between mean and median for the variable “age” are not statistically
significant.
The results for the variable BVE, due to the CSR principle, indicate that the
value of assets increases, in order to enhance future results. However, and as
expected, the increase in the explanatory power of the model is very low (3.41% in
the sample of net firms and 3.08% in the sample of non-net firms), which confirms
that the main determinant of the value of these companies is the persistence of the
variable “results”.
To examine the impact of the variable advertising, the model 6.7 was estimated.
The results are shown in Tables 7.9 and 7.10.
For the sample of net firms, we could not estimate the model for the group of
companies with profits, given the small number of observations. For the group with
losses, we only have information since the year 1998. Note that the boom of
observations focused on the period dot.com bubble, i.e. 1999–2000.
Regarding the results obtained, these are similar to those obtained for the
B2B_R&D group. The variable advertising has a positive and statistically signifi-
cant coefficient but only at a 10% level, indicating that the market seems to value
Year # Na Sample with losses
# Nb ac0 BVEc Res_Advc Advc Adj. R2
(%)
1998 7 18 4.28*** 0.006*** −0.003 0.06*** 56.64
(0) (11.377) (4.756) (−1.229) (5.033)
1999 16 92 5.28*** 0.004*** −0.002 0.008 27.04
(2) (5) (27.533) (4.352) (−0.932) (0.83)
2000 15 110 3.15*** 0.001** 0.0004 0.018*** 27.23
(1) (2) (16.983) (2.289) (1.398) (4.673)
2001 0 78 2.85*** 0.003*** 0.0021** 0.019*** 42.54
(4) (12.123) (3.598) (2.046) (3.221)
2002 10 48 2.75*** 0.004*** 0.003 0.012 40.10
(2) (4) (9.764) (4.251) (0.943) (1.354)
2003 13 23 3.41*** −0.0002 −0.013 0.038*** 28.47
(1) (2) (8.377) (−0.0002) (−1.927) (3.344)
Meand 3.618*** 0.003* −0.002 0.026* 37.00
T-statisticse 5.417 3.218 −0.837 3.07
Estimated model
MVEit = a0 + a1 (BVE)it + a2 (Res_Adv)it + a3 (Adv)it + eit
where MVE represents the market value of the company, BVE the book value of equity, Res_Adv
the results before extraordinary items and Adv the amount invested in advertising
a
Insufficient number of observations to estimate the regression model 4.7 for the subsample with
profits
b
c
d
e
The T-statistic (two-tailed test), calculated from the ratio of the average standard deviation
multiplied by 6½
(***), (**) and (*) indicates significance level of 1, 5 and 10%, respectively
this variable as an asset and not a cost. The variable results despite the adjustment
investment of the investment in advertising remain negative, but not statistically
significant. Note that the results for this group should be treated with some caution,
given the high volatility of the number of observations. In 2003 we only have 23
companies, while in 2000 the number of these companies was 110.
For the sample of non-net firms, it is also impossible to estimate the model 6.7
before the year 1999 due to the small number of observations. Regarding the results
for both groups (companies with profits and companies with losses), the findings
again show an asymmetry in the market in assessing the variable “advertising. This
variable reports a negative (positive) coefficient significant at level a 5% (1%) in
sample with profit (losses). Hence, the results supported Hypotheses 3a and 3b.7
7
Remember that Hypothesis 3a is post-dated: There is a positive relationship between the variables
R&D and advertising and MVE in companies in the start-up phase, reporting losses. Hypothesis
3b predicted the inverse relationship for profitable firms in a more mature phase.
154
# ab0 BVEb Res_Advb Advb Adj. # a0b BVEb Res_Advb Advb Adj.
Na R2 Na R2
(%) (%)
1999 25 4.01*** −0.0001 0.08*** −0.056 34.4 19 2.92*** 0.006*** −0.004 0.071 53.82
(1) (8.767) (−0.016) (4.0) (−1.07) (0) (4.285) (4.426) (−0.885) (1.074)
2000 22 3.11*** −0.001 0.13*** −0.151*** 72.21 36 2.76*** 0.003*** −0.001 0.055** 39.25
(1) (9.844) (−0.032) (6.853) (−5.04) (1) (8.104) (3.854) (−0.515) (2.305)
2001 81 3.80*** 0.005*** 0.052*** −0.052** 52.94 79 3.18*** 0.003*** 0.001 0.09*** 34.20
(3) (22.965) (3.79) (5.359) (−2.285) (4) (13.981) (2.711) (1.077) (2.963)
2002 89 3.77*** 0.003*** 0.039*** −0.016 52.79 65 2.55*** 0.007*** 0.009*** 0.071** 34.96
(2) (23.703) (4.015) (5.077) (−0.819) (5) (8.73) (4.105) (3.085) (2.408)
2003 86 4.49*** 0.002*** 0.021*** −0.021*** 52.43 51 3.48*** 0.003*** −0.004 0.03 30.53
(1) (31.428) (4.055) (4.9) (−4.852) (7) (12.614) (2.368) (−0.672) (1.005)
Meanc 3.835*** 0.002** 0.065** −0.059** 52.95 2.98*** 0.004*** 0.0003 0.063*** 38.55
d
T-statistics 21.778 2.242 4.3 −3.076 23.228 6.365 0.177 8.019
Estimated model
MVEit = a0 + a1 (BVE)it + a2 (Res_Adv)it + a3 (Adv)it + eit
where MVE represents the market value of the company, BVE the book value of equity, Res_Adv the results before extraordinary items and Adv the amount
invested in advertising
a
Number of observations with negative value for the variable BVE in brackets
b
c
d
The T-statistic (two-tailed test), calculated from the ratio of the average standard deviation multiplied by 5½
(***), (**) and (*) indicate a significance level of 1, 5 and 10%, respectively
Given that this research proposes as a proxy to the phase of the life cycle of the
company the variable results, i.e. companies with profits are assumed to be
undergoing a steady growth/maturity phase compared to firms with losses in the
start-up phase, the previous results show that the market assesses asymmetrically
the investments in R&D and advertising, according to whether companies report
profits or losses, as documented in Tables 7.7, 7.8, 7.9 and 7.10. Therefore, we
conclude that the market evaluates differently the determinants of the value of the
company/industry as it moves toward maturity.
These results find theoretical support in Proposition 9 of the FOM, and due to the
conservatism accounting effect, i.e. in the early years of the life of a company, the
company reports negative results, because only a small portion of investment is
capitalized and the remainder is immediately considered as a cost (e.g. investment
in R&D and advertising) which reduces the cash flows of the period. However,
given the principle of rationality, the company only continues to invest if the
investment opportunities generate abnormal returns. Moreover, these results are
consistent with those obtained by Sougiannis (1994), who documents the duality of
the effect of the short and medium/term associated with investments in R&D,
whereas investors associated a positive informational content with current invest-
ments in R&D and advertising (direct effect). However, the indirect effect resulting
from the capitalization of income generated by the investments in the past is more
statistically significant.
Hypothesis 5 tests empirically whether the variations in the MVE in net firms and
non-net firms are explained by the same determinants, i.e. results, BVE, R&D and
advertising. Although the results obtained are very similar for the two samples (see
Tables 7.1, 7.2, 7.3, 7.4, 7.5, 7.6, 7.7, 7.8, 7.9 and 7.10), the results in Table 7.1
show that the effect of a “positive valuation of losses” is more pronounced in net
firms in line with the results obtained by Lee (2001) and Cooper et al. (2001). These
authors document that, with the simple inclusion of the term “dot.com” in the
company name, sometimes without any substantial change in the core business of
the company, the companies showed significant abnormal returns, which tended to
persist over time. Also Bartov et al. (2002), when analyzing the process of the IPO
pricing, noted that the variables “losses” and “negative cash flow” are positively
valued by the market with reference to net firms as compared to a sample of
contemporaneous IPOs. In line with these results, we associate a fashionable effect
(fad) with the sample of net firms, which relies, in our opinion, on the emerging
nature of this sector, and the strong expectation associated with this sector in the
near future.
References
Bartov E, Mohanram P, Seethamraju C (2002) Valuation of internet stocks—an IPO perspective.

J Account Res 40(2):321–346
Basu S (1997) The conservatism principle and the asymmetric timeliness of earnings. J Account
Econ 24:3–37
Ben-Zion U (1978) The investment aspect of nonproduction expenditures: an empirical test.
J Econ Bus 30:224–229
values over the past forty years. J Account Econ 24(1):39–67
of equity. Account Rev 74(1):29–61
Cooper MJ, Dimitrov O, Rau R (2001) A Rose.com by any other name. J Finance 56(6):2371–
2388
of what has changed? J Account Econ 34(1 & 3):43–67
Dechow P, Hutton AP, Sloan RG (1999) An empirical assessment of the residual income valuation
model. J Account Econ 26:1–34
Demers E, Lev B (2001) A rude awakening: the internet shakeout in 2000. Rev Account Stud 60
(2&3):331–359
Fama E, MacBeth J (1973) Risk, return, and equilibrium: empirical tests. J Polit Econ 81:607–636
activities. Contemp Account Res 11(2):689–731
Frazen LA, Radhakrishnan S (2009) The value relevance of R&D across profit and loss firms.
J Account Public Policy 28(1):16–32
Greene WH (2000) Econometric analysis, 4th edn. Prentice Hall, New York
Gujarati DN (2002) Basic econometrics, 4th edn. McGraw-Hill, New York
(eds) Intangibles: an Oxford Management Reader. Oxford Press University, Oxford
Jorion P, Talmor E (2006) Value relevance and non financial information in emerging industries:
the changing role of web traffic data. Working Paper, University of California: Irvine e London
Business School. http://ssrn.com/abstract=258869
VDM
Lee PM (2001) What’s in a name.com? The effects of “.com” name changes on stock prices and
Ljung GM, Box GP (1978) On a measure of lack of fit in time series models. Biometrica 66:66–72
pricing. Working Paper, Polytechnic University of Valência—Department of Economics and
Social Sciences. http://dx.doi.org/10.2139/ssrn.328781
(4):411–433
Newey W, West K (1987) A simple, positive semi-definite heteroscedasticity and autocorrelation
consistent covariance matrix. Econometrica 55:703–708
Sougiannis T (1994) The accounting based valuation of corporate R&D. Account Rev 69(1):44–68
References 157
Commer 4(4):23–40. Doi:10.1300/J179v04n04_02
stocks. J Account Res 38(Supplement):137–162
White H (1980) A heteroscedasticity consistent covariance matrix estimator and direct test for
heteroscedasticity. Econometrica 48:817–838
Part IV
Conclusions
Chapter 8
Conclusions and Suggestions for Further
Research
Abstract This chapter presents the main conclusions of the study and suggestions
for further researcher. Based on the results obtained, we provide also some rec-
ommendations to policy makers, financial analysts and regulators.
Keywords Persistence of losses

Positive valuation of losses Ohlson and

Feltham and Ohlson models Net firms
The central objective of this research was to analyze the apparent anomaly between
the report of losses and the high market value of equity registered by companies in
emerging sectors. This study examines a specific set of companies—new US
companies—net firms. This effect is not totally new: for instance, Kothari and
Zimmerman (1995: 176) had identified this phenomenon; however they failed to
provide any explanation for it. Thus, given the magnitude of this phenomenon in
the 90s, our research explains this phenomenon as:
(i) Due the accounting conservatism effect, the information reported by the
financial statements underestimate the growth opportunities owned by these
companies. For example, Trueman et al. (2000, 2001), Hand (2001),
Martínez and Clement (2002), and Rajgopal et al. (2002), based on the work
of Amir and Lev (1996), who for the first time introduced non-financial
variables into the valuation models, demonstrate that the weak explanatory
power of financial variables is partly offset by the inclusion of web traffic
variables, which better capture the value chain of these companies, facili-
tating the prediction of future profitability, especially the volume of sales;
(ii) When we analyzed the life cycle (eight years) and used the methodology used
by Fama and MacBeth, we aim to enhance the relevance over time of the
determinants of value for these companies. Loughran and Ritter (2003) and
Ljungqvist and Wilhelm (2003), in line with current literature, demonstrated
that IPOs tend to have a cluster effect in time. Knauff and van der Goot (2001)
and Bartov et al. (2002) confirm this effect in this group of companies.
(iii) Given the investment profile that characterizes this type of company,
investment in intangible assets represent a “slice” of the significant value

DOI 10.1007/978-981-10-3009-3_8
162 8 Conclusions and Suggestions for Further Research
allocated to the expectations of a higher probability of the existence of

growth opportunities, hence the greater the risk of such securities (Myers
1977). These results complement the results of Rajgopal et al. (2002) who
document the persistence of abnormal returns associated with these securi-
ties. Ofek and Richardson (2002, 2003) justify these abnormal returns as a
reward against the higher level of risk inherent to them. The results are
robust, given the representativeness of the sample, the reporting period (eight
years), the accuracy in controlling the survival effect (survivor bias effect), by
the inclusion of companies with negative equity (against the current practice
in empirical studies) and the comparison of the results with a control sample
—non-net firms (match sample).
Thus, the key findings are:
(i) Companies created during the period of the New Economy Period (NEP)—
net firms and others (e.g. contemporaneous IPOs) are mostly
technology-based, and have a focused investment profile in intangible assets;
the increase in sales tends to be accompanied by increased investment in
R&D and advertising, even with the increasing age of the company, which
explains that the group with losses is the most numerous and persistent group
of companies. This group is the group that invests more aggressively in
intangibles as a proportion of their sales volume. These results confirm that
losses in companies in start-up phase/growth, particularly those that are
technology-based, are a result of accounting conservatism effect, as modelled
by the FOM, for accounting purposes and in compliance with GAAP.
(ii) When assessing how the effect of “conservatism accounting” affects the rela-
tionship between the market value of the equity of these companies and the
results—losses, we conclude that the market positively evaluates the variables
R&D and advertising in the group with losses in both samples net and non-net
firms, confirming the opportunity investment hypothesis. Investors seem not to
fix their attention on the variable results as an aggregate variable, but associated
a higher volume of investment to R&D and advertising to the probability of the
existence of larger growth opportunities. This result confirms the clientele
effect, documented by Chan et al. (2001), characteristic of this type of com-
pany. The persistence of the investment in these items over the reporting period
(eight years) reveals the confidence of managers in the projects in their port-
folio. For the group with profits, these variables (i.e. R&D and advertising) are
valued negatively, i.e. as costs, predominant the indirect effect (Sougiannis
1994), i.e. the variable results already reflect the effect of the investments made
in the past. Thus, as demonstrated by Modigliani and Miller (1966), the results
are the main determinant of the value of these companies; thus the statistical
significance of the variable results increases with its persistence.
(iii) The statistical relevance of the phenomenon of “positive valuation of the
losses” is large in the sample of net firms. This result can be explained by the
following facts:
8 Conclusions and Suggestions for Further Research 163
– The group of net firms includes a larger number of companies with high
losses. Even after the adjustment of the variable results by the invest-
ments in R&D and advertising, its value remains negative, showing an
effect of conservatism accounting which is more severe in this sample.
This result indicates the need to include other variables for calculating the
growth opportunities, such as proxies for web traffic;
– This group contains a higher percentage of technology-based companies
and;
– The volatility is higher, particularly in the subgroup with losses in con-
sequence of the strong uncertainty that characterize the investments in
intangible assets (Kothari et al. 2002).
(iv) However, the results of Collins et al. (1999), who argue that the phenomenon
of positive valuation of losses is due the omission of the variable equity
(BVE) from the model, are partially confirmed. With the introduction of the
variable equity (BVE) in the valuation model, the phenomenon of positive
valuation of losses disappears in the sample of non-net firms, persisting but
with no statistical significance in the sample of the net firms. The importance
of including this variable in the valuation models is unquestionable for the
group with losses in both samples. The inclusion of this variable in the
valuation models registers an increase explanatory power in both samples.
This result shows and, in line with OM and FOM models, that variable
equity (BVE) is a proxy for expected future abnormal results, given the
limited information content of the variable results when the company report
losses. In this context, the theory of abandonment option, which associates a
higher probability of liquidation to a company that show a higher persistence
of the losses, seems not to be appropriate. Thus, the report of loss cannot be
indicative of a process value destruction.
(v) We emphasize also that the BVE variable is assumed as a tool to reduce
agency costs, particularly with creditors as it identifies itself as a proxy for
the “recognized assets”, given the predominance of intangible-based assets.
In summary, this study focused on the relationship between market capitalization
and profitability (negative) of the companies in the “new economy”, in US net firms.
Although not entirely new, this phenomenon will certainly be repeated although
with some specific characteristics. So given the results obtained, we highlight the
main contributions of this research:
(i) Increasing investment in the 90s, and in line with McCallig (2004) and Joos
and Plesko (2005), appears linked to a change in the business profile of firms
operating in the market: small companies, mostly technology-based that
report a higher magnitude losses for longer periods;
(ii) The report of losses may not be indicative of value destruction, in clear
opposition to the theory of abandonment option;
(iii) Therefore, the information content of the losses is not irrelevant for assess-
ment purposes, if they arise in association with them, particularly in the
start-up phase/growth performance of high growth opportunities;
(iv) Given the asymmetrical nature by the market in the assessment of the
variables R&D and advertising, and hence the variation of statistical sig-
nificance of the variable results over time, we find that the Internet industry is
still an emerging sector, associated with new business opportunities, so it is
wrong to treat equally all companies that report losses, as this may lead to
erroneous empirical findings;
(v) Companies in financial stress, particularly technology-based, tend to opt for
mergers and acquisitions (M&A) processes as a restructuring strategy to the
detriment of bankruptcy, which would imply a greater destruction of value.
Based on these results, some recommendations are relevant to policy makers,
financial analysts and regulators:
(i) agents (managers) should define strategies to generate high cash flows and
appropriate economic rents, avoiding the pitfalls of a fashion phenomenon
and situations of capital myopia. As specified by the models of OM and
FOM, the abnormal returns tend to converge quickly to the industry average
due to the actions of the competition.
(ii) Financial analysts should be based on the potential for profit generation and
its growth in order to avoid situations of overvaluation and the generation of
financial bubbles;
(iii) Regulatory authorities should review the rules of reporting and publishing of
financial information to provide investors a broader intelligence picture and a
more timely window to make their investment decisions.
This work also raises a set of lines for future research:
(i) It is important to extend the review period towards a better assessment of the
interaction between the effect of the life cycle and the value created by
continuous investments in R&D and advertising.
(ii) Another extension of this research would be a comparative study with the
traditional sectors, in order to analyze potential differences and similarities.
For example: what is the impact of the Internet on generation of sales, on the
structure cost and on the creation of new business opportunities or the
internationalization process.
(iii) The characteristics of the random term OM model give the investigator the
freedom to define the functional form of the model to make estimates, which
depends on assumptions about the relationship between the dependent
variables, independent, and the random term. Given the results of the
Ramsey test are systematically significant, this indicates an incorrect
potential model specification; moreover, the results obtained under Eq. 7.11
8 Conclusions and Suggestions for Further Research 165
reveal a possible existence of non-linear relationships between variables (Ye

and Finn 1999). Thus, another potential extension of this investigation would
assume that the random term is multiplicative instead of additive. In this
context, Box–Cox transformation would prove to be more suitable.
(iv) Given the suggestion of Ohlson (2000), and having available the analysts’
forecasts of future results, it would be interesting to use this variable as a
proxy for the vt variable, i.e. other non-financial information. In addition, it is
relevant to validation if the abnormal results follow an autoregressive first
order process AR (1). For example, Callen and Morel (2001) show that the
abnormal results tend to follow an autoregressive process of second-order
AR (2) and not of the first-order AR (1), as predicted by OM and FOM
models.
From the modelling point of view, the challenges are equally large. It is relevant
to introduce more flexibility to the assumptions of OM and FOM, e.g. the prefer-
ences (beliefs) of the investors are not homogeneous and the capital markets are not
perfect; thus it is important to introduce the effects of asymmetric information,
transaction, bankruptcy and agency costs, and the taxes effect. Furthermore, some
studies show non-linear relationships between the variables: “market capitaliza-
tion”, “equity” and “results”, particularly when this variable reaches extremes. For
example, Yee (2000) and Zhang (2000) find that whenever the company adjusts its
investment policy in terms of results, this generates a non-linear relationship
between the market value of equity and the company’s results. Ye and Fin (1999)
show that when the abnormal rate of return on equity (not the abnormal results)
follows an autoregressive first order process and the company does not pay divi-
dends (assume therefore that the company is still in the start phase-up/growth), then
the company’s value is not a linear function of the equity and results. Burgstahler
and Dichev (1997) also conclude that there is a convex relationship between the
market value of equity and the company’s results, when they reach unsatisfactory
levels.
Thus the determination of the company’s value due to accounting financial
variables in a non-linear relationship presents a high potential for future research.
Bernard (1995: 735) noted:
The Ohlson model represents the base of a branch (for) capital market research … Ohlson
(1995) and Feltham and Ohlson (1995) return to” step one “and attempt to build a more
solid foundation for further work. Our challenge is clear.
The determination of the sources of enterprise value is a central concern of

business finance which is far from exhausted, even in traditional sectors, with
greater stability and less contestability. By analyzing the most dynamic industrial
sector in recent years, we hope to have contributed to the clarification of this issue
by introducing the theoretical framework of the OM and FOM models which have
scarcely been used in empirical studies.
References

industry. J Acc Econ 22(1&3):3–30
Bartov E, Mohanram P, Eethamraju C (2002) Valuation of internet stocks—an IPO perspective.
J Acc Res 40(2):321–346
Bernard VL (1995) The Feltham-Ohlson framework: implications for empiricists. Contemp Acc
Res 11(2):733–747
Burgstahler DC, Dichev I (1997) Earnings, adaptation and equity value. Acc Rev 72(2):187–215
Callen J, Morel M (2001) Linear accounting valuation when abnormal earnings are AR (2). Rev
Quant Financ Acc 16(3):191–203
activities.Contemp Acc Res 11(2):689–731
Joos P, Plesko GA (2005) Valuing loss firms. Acc Rev 80(3):847–870
Knauff P, van der Goot T (2001) The relevance of reported financial information for valuing
European IPOs, EFMA Lugano Meetings. http://ssrn.com/abstract=263176
Kothari SP, Zimmerman JL (1995) Price and return models. J Acc Econ 20:155–192
Kothari SP, Laguerre TE, Leone AJ (2002) Capitalization versus expensing: evidence on the
uncertainty of futureearnings from capital expenditures versus R&D outlays. Rev Acc Stud 7
(4):355–382
Ljungqvist A, Wilhelm WJ (2003) IPO pricing in the dot. com bubble. J Financ 58(2):723–752
Loughran T, Ritter J (2003) Why has IPO underpricing changed over time. Working paper
presented at AFS—American Finance Association 2003, Washington Meeting
pricing. Working paper, Polytechnic University of Valência—Department of Economics and
firms. Paper presented at 14th annual conference on Financial Economics and Accounting
Modigliani F, Miller M (1966) Some estimates of the cost of capital to the electric utility industry
1954–1957. Am Econ Rev 56(3):333–391
Ofek E, Richardson M (2002) The valuation and market rationality of internet stock prices. Oxf
Rev of Econ Policy 18(3):265–287
Ofek E, Richardson M (2003) Dot. com mania: the rise and fall of internet stocks prices. J Financ
58(3):1113–1138
(2):661–687
Ohlson J (2000) Earnings, book values, and dividends in equity valuation: an empirical
perspective. Contemp Acc Rev 18(1):107–120
Rajgopal S, Venkatachalam M, Kotha S (2002) Managerial actions, stock returns, and earnings:
the case of business-to-business internet firms. J Acc Res 40(2):529–556
Sougiannis T (1994) the accounting based valuation of corporate R&D. Acc Rev 69(1):44–68
stocks. J Acc Res 38(Suppl):137–162
References 167
Trueman B, Wong MHF, Zhang X (2001) Back to basics: forecasting the revenues of internet
firms. Rev Acc Stud 6:305–329
Ye J, Finn M (1999) Nonlinear and nonparametric accounting-based equity valuation models.
Working Paper, Baruch College, University of New York, New York
Yee KK (2000) Opportunities knocking: residual income valuation of an adaptive firm. J Acc
Audit Financ 15(3):225–266

Equity Valuation and Negative Earnings - The Case of The Dot - Com Bubble (PDFDrive)

Uploaded by

Copyright:

Available Formats

Equity Valuation and Negative Earnings - The Case of The Dot - Com Bubble (PDFDrive)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Equity Valuation and Negative Earnings - The Case of The Dot - Com Bubble (PDFDrive)

Uploaded by

Copyright:

Available Formats

What is the topic of research covered in the document?

What is the topic of research covered in the document?

What themes are featured in volumes published in the series?

What themes are featured in volumes published in the series?

Accounting, Finance, Sustainability, Governance & Fraud:

Theory and Application

Ana Paula Matias Gama

More information about this series at http://www.springer.com/series/13615

Marco Antonio Figueiredo Milani Filho

Liliane Cristina Segura

ISSN 2509-7873 ISSN 2509-7881 (electronic)

© Springer Nature Singapore Pte Ltd. 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

ISCTE-IUL José Paulo Esperança

Part II Literature Review

4 The Impact of Investment in Intangible Assets

Part III Empirical Study

Table 3.1 Summary of the factors supporting the weak correlation

Table 7.3 Correlation matrix—sample of net ﬁrms . . . . . . . . . . . . . . . . . . . 144

1.1 Initial Comments

© Springer Nature Singapore Pte Ltd. 2017 3

1.2 The Internet: History and Concepts

1.3 The Internet and Electronic Commerce

Due to new technologies, business have adapted to available tools, targeting

1.4 e-Business Environment

Internet-technology-oriented business (e-business) improves productivity and

Business Intelligence (BI): is the production of management information to support

1.5 Dot.com Companies

The NASDAQ (National Association of Securities Dealers Automated Quotation)

1.7 Economic Bubbles

An economic or speculative bubble could be described as a situation in which asset

as herd behaviour. Banerjee (1992) deﬁnes herding as “everyone doing what

1.8 The Valuation of Internet Companies

revenues in conjunction with adjustments in operating margins; and reduction of

It must be ascertained whether the phenomenon is transient, abnormal, or if the

Accepted Accounting Principles (GAAP), investments in intangible assets are

1.9 Objectives and Main Research Contributions

of the existence in the portfolio of greater opportunities for growth, so higher

1.10 Organization of the Work

We have structured the present book therefore as follows.

Amir L, Lev B (1996) Value-relevance of nonﬁnancial information: the wireless communications

Abstract This chapter analyses the phenomenon of “positive valuation of losses” in

Keywords Technology-based companies Positive valuation of losses Ohlson

2.1 The Ohlson Model (OM)

© Springer Nature Singapore Pte Ltd. 2017 19

bvt ¼ bvt1 þ xt dt ð2:2Þ

@bvt = @dt ¼ 1 ð2:3aÞ

@xt = @dt ¼ 0 ð2:3bÞ

xat ¼ xt ðRf 1Þbvt1 ð2:4Þ

This expression, after some algebraic transformations, allows us to redeﬁne the

Pt ¼ bvt þ a1 xat þ a2 vt ð2:8Þ

Pt ¼ j½uxt dt  þ ð1 jÞbvt þ a2 vt ð2:9Þ

j ¼ ðRf 1Þa1 ¼ ðRRff1 Þw

The previous expression can be interpreted as a weighted average of the eval-

2.2 The Extent of the Ohlson Model: The Feltham Ohlson

In this context the variables considered are:

(ii) Net interest relation (NIR)—Income generated from non-operating activities:

it ¼ ðRf 1Þ fat1 where Rf ¼ 1 þ rf : ð2:12Þ

Pt ¼ j½uxt dt þ ð1 jÞbvt þ a2 vt ð2:9Þ

fat ¼ fat1 þ it ½dt ct : ð2:13Þ

Pt ¼ bvt ½1 a1 ðRf 1Þ þ a2 vt þ a1 xt þ a1 ðRf 1Þxt a1 ðRf 1Þdt and j ¼