Anatomy of Stock Market Bubbles

Anatomy of Stock Market Bubbles
ICFAI Books
An Introduction
ICFAI BOOKS is the initiative of the ICFAI University Press to publish

a series of books in the areas of finance, management and allied
areas with a special focus on emerging and frontier areas. These
books seek to provide, at one place, a retrospective as well as
prospective view of the contemporary developments in the
environment, with emphasis on general and specialized branches of
knowledge and applications.
The articles are organized in a sequential and logical way that

makes reading continuous and helps the reader acquire a holistic
view of the subject. This helps in strengthening the understanding of
the subject better and also enables the readers stretch their thoughts
beyond the content of the book. The series is designed to meet the
requirements of executives, research scholars, academicians and
students of professional programs. The ICFAI University Press has
published over 750 books in this series. For full details, readers are
invited to visit our website: www.icfaipress.org/books.
Anatomy of
Stock Market Bubbles
György Komáromi
ICFAI BOOKS
The ICFAI University Press
ANATOMY OF STOCK MARKET BUBBLES
Author: György Komáromi
Copyright © György Komáromi, 2006. All rights reserved.

Although every care has been taken to avoid errors and omissions, this publication is being
sold on the condition and understanding that the information given in this book is merely
for reference and must not be taken as having authority of or binding in any way on the
author, editor, publisher or sellers.
Neither this book nor any part of it may be reproduced or transmitted in any form
or by any means, electronic or mechanical, including photocopying, microfilming
and recording or by any information storage or retrieval system, without prior
permission in writing from the copyright holder.
Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation without
intent to infringe.
Only the publishers can export this book from India. Infringement of this condition
of sale will lead to civil and criminal prosecution.
First Edition: 2006
Printed in India
Published by The ICFAI University Press

52, Nagarjuna Hills, Punjagutta
Hyderabad, India–500 082
Phone: (+91) (040) 23430 – 368, 369, 370, 372, 373, 374
Fax: (+91) (040) 23352521, 23435386
E-mail: [email protected], [email protected], [email protected]
www.icfaipress.org/books
ICFAI Editorial Team:

Editorial Co-ordinator : K Krishna Chaitanya
Sub-editor : K Prabhakar and Aditya Ghosh
Cover Designer : T Bal Reddy
Designer : D S Bhushan Raj
ISBN: 81-314-0408-0
Dedication
To my wife and my daughter
About the Author
Dr. György Komáromi is Assistant Professor of Finance at CEU

Business School in Budapest and College Professor in Finance at
International Business School, Hungary. He teaches different courses
in Finance in under- and postgraduate programs. He is currently
working on the research project titled “Economic role and importance
of the stock exchange in Hungary between 1864 and 2005” supported
by OTKA (Hungarian Scientific Research Fund). His research
interests also include behavioral and experimental aspects of
financial decisions. He was educated at University of Miskolc
(Hungary), Graduate School of Management of Rouen (France), and
Pannon University (Hungary), where he received his PhD in
Management and Business Studies in 2005. He has been full-time
Assistant Professor in Finance at Management Development Centre
of Budapest Corvinus University (1999-2001), and at Pannon
University (2001-2006). He taught as a visiting lecturer at Budapest
University of Technology and Economics (2004-2005) and has been
ERASMUS exchange professor at University of Paris and Rovaniemi
Politechnique, Finland (2002). He published several articles on topics
such as Behavioral Finance, Corporate Finance and Institutional
Economics and was awarded the Earhart Scholarship (USA) in 2003.
He is a regular participant in international conferences and
workshops.
Acknowledgments
First of all I am grateful to Ákos Szegó, who provided me a great help in
translating my book, and provided critiques and comments. I am also
indebted to Péter Mihályi, Balázs Hámori, Katalin Szabó and András Vigvári
for their great comments and suggestions. I acknowledge Iván Major,
Philip Keefer, John Nye, and Lee Benham for the inspiring discussions
that made my work better.
I am also thankful to my colleagues at Pannon University and the

participants in the Workshop on Institutional Analysis organized by Ronald
Coase Institute in 2003. I was also benefited by the international conferences
organized by INFIDENT Scientific School in Debrecen between 2002 and
2004. I would like to express my heart-felt thanks to OTKA (Hungarian
Scientific Research Fund) for supporting my research project (F60619).
I am also very thankful to Krishna Chaitanya for giving me an important

assistance during the publication process.
I am grateful to my wife, my daughter and my family for their patience and

support when I was working on this book.
György Komáromi.
Contents
List of Figures and Tables I

Preface III
1. Different Aspects of Bubbles 1

1.1 Asset Price Bubbles in a Mathematical Approach 2
1.1.1 Rational versus Speculative Bubbles 3
1.1.2 The Problem of Determining
Fundamental Value 9
1.1.3 Behavioral Finance on Bubbles 15
1.1.4 The Predictability of Crash 21
1.1.5 Conclusions 22
1.2 Asset Price Bubbles in Laboratory 24
1.3 Literary Economics and Bubbles 30
1.3.1 A Classic Example: Dutch Tulip Bulbs 32
1.3.2 The First Stock Market Bubbles:
Mississippi and South Sea Companies 38
1.3.3 The Crashes of 1929 and 1987 45
1.3.4 Internet Bubble 53
2. Anatomy of Stock Market Bubbles 69

2.1 Investment Decisions and Noise Trading 69
2.1.1 Illusion of Knowledge in
Investment Decisions 73
2.1.2 Co-movement of Stock Prices 74
2.2 How to Distinguish Stock Market Bubbles? 79
3. The Hungarian Case, 1996-2003 88
4. Summary 95
References 103
Index 111
List of Figures and Tables
I. Different Aspects of Bubbles 1
Figure 1.1: Stock Value if There are no Trading Opportunities 10
Figure 1.2: Stock Value if There are Trading Opportunities 11
Table 1.1: Main Behavioral Patterns in Investors Decisions 18
Table 1.2: Factors Determining the Extent of Price Bubbles
in Laboratory Experiments 27
Figure 1.3: DJIA, NASDAQ and DOT between 1995 and 2001 56
Figure 1.4: Changes in Share Possession of US Households 57
2. Anatomy of Stock Market Bubbles 69

Figure 2.1: Inference Problem from Price Changes 70
Figure 2.2: Daily Share Co-Movement in the Budapest
Stock Exchange 78
Table 2.1: (I.) Characteristics of Stock Market Bubbles in
Economic History 83
Table 2.1: (II.) Characteristics of Stock Market Bubbles in
Economic History 84
Table 2.1: (III.) Characteristics of Stock Market Bubbles in
Economic History 86
3. The Hungarian Case, 1996-2003 88

Figure 3.1: BUX Between January 1996 and September 2003 89
Figure 3.2: Trade Volume and Turnover Velocity Between
1996 and 2003 90
Figure 3.3: Structure of Share Ownership in BSE Between
1997 and 2003 91
Figure 3.4: Daily Co-Movement of BUX and Shares
Between 1997 and 2003 92
Preface
“We know there are bubbles but we cannot even demonstrate

their existence (our indicators are not up to the task). We know
that given specific circumstances, bubbles «burst» but we cannot
yet define necessary and sufficient conditions.”
Csontos L., Király J. and László G. (1999. p.586).
Economists often use the term “stock market bubble” to back up

their market analysis or explain past and present events. It is part
of the professional jargon, one of the colorful, emotionally charged
expressions used in economics. But what exactly is a bubble?
I have first heard about the notion during my university studies,

mentioned by a professor in a course about capital markets. The
spectacular rise of the Hungarian stock exchange during the 1990s
fascinated many a university student (including me), who became
personally involved as well, standing in queue for shares of
privatized state companies. At the time, technical issues regarding
IV
the stock exchange looked more relevant than any financial theory
because seemingly everybody could profit from buying shares.
Stock market analysis became part of the daily routine.
I have considered claims like “shares are considerably overvalued

and a market correction is inevitable” to be superficial and far
from the realm of real finances; explanations like “prices are blown
like bubbles” seemed empty to me. The theory of efficient markets,
based on the simplifying but apparently valid presumption of
rational investors, was a milestone inasmuch as it carried the
message that any future return on my investment in shares was
only a matter of luck. I have also realized that “the starting point
[of pricing financial assets] is the fact that we cannot beat the
market using any mathematical model” (Medvegyev 2002. p.597),
and I grew skeptic towards those who attributed their stock market
profits to their personal abilities.
I was impressed by Burton Malkiel’s book, Random Walk Down

Wall Street (1990), which reinforced my belief that “there is no
such thing as a stock market guru” and future returns of shares are
basically unpredictable. But additional questions kept coming up.
How and why do large-scale stock exchange hausses and baisses
form? Are there always economic reasons behind them, or could
there possibly be some other explanations?
During my university years there were three books contributing

to an impression I had: “strict” financial theories were skirting the
important question of what causes a wild fluctuation and eventual
collapse of share prices? André Kostolany’s book Kostolanys
Börsenpsychologie (1991) emphasized the irrationality of human
behavior. During our long telephone conversations he himself
made the point that “he who is studying modern economics will
never understand the true driving force behind share price
V
fluctuations”. John Kenneth Galbraith’s satirical essay A Short

History of Financial Euphoria (1994) supplied historical examples
for the recurring and predictable commonalities arising from folly
and complacency. Still, it was the chapter titled The State of Long-
term Expectations in John Maynard Keynes’ General Theory (1936)
that provided me with a scientific foundation for the „workings”
of capital markets as well as for investor psychology and behavior,
having the most significant impact on my thinking, inspiring the
commencement of writing of this work. I will continue to refer
to these books in the following chapters.
The scientific question comes from the opening quote: how

and why do price bubbles form and burst, and what are the
necessary and sufficient conditions? There are several studies in
the literature dealing with these questions, and from the 1990s
on we could come across an increasing number of scientific and
pseudo-scientific articles and studies on stock market price changes.
Examples of the former: Brunnermeier (2001), Shiller (2000),
and articles in the 1990/4 issue of the Journal of Economic
Perspectives. The contradicting results and not easily substantiable
conclusions of a wide-ranging research activity even cast a doubt
on whether bubbles actually exist at all. This suggests that
economists are using the concept for lack of anything better as
they cannot make scientifically valid claims about the underlying
events.
My objective is to point out weaknesses in the main theories

and I also attempt to integrate the interpretation of stock market
bubbles into a valid conceptual framework. This essay based on
my dissertation (Komáromi 2004) seeks to provide an answer to
when we can term a stock market phenomenon as a bubble.
VI
The Subject and Structure of the Book

There exists no uniform economic theory to explain stock market
bubbles. The first chapter of the essay outlines competing
explanations which can be classified into three main schools of
thought: the mathematical and empirical approach and that of
literary economics.
Studies classified by this distinction as mathematical use the

rational equilibrium price as their starting platform. The first
subgroup consists of studies that analyze share price development
in the presence of symmetric and asymmetric information
assuming rational expectations. We will take closer look at the
most important conclusions of these models, like the issue of
fundamental value or “herding”. The other subgroup is the field
of behavioral finance, which is based on psychological
characteristics supported by the findings of empirical research and
observable during investment decision making. We will examine
these behavioral patterns and the role they play in triggering price
changes, also commenting on the predictability of the collapse of
bubbles. Further clarification is needed because necessary and
sufficient conditions for the formation and bursting of stock
exchange bubbles, also useful in practice, cannot be provided
within the framework of mathematical economics.
After this we will present the achievements of empirical

economics. Stock exchange simulations carried out in laboratories
support some conclusions of the previously discussed theories,
but they also provide further insight regarding the phenomenon.
When assessing bubbles examined in documented laboratory
experiments we face the problem of not knowing exactly to what
degree experimental circumstances distort results.
VII
Although literary economics of a mostly historical-logical

approach does not belong to the arsenal of mainstream economics,
for better understanding of bubbles a recapitulation and analysis
of historical financial case studies and intuitive examples and
conclusions is evidently necessary. We will analyze the “Dutch tulip
mania”, the stories of the Mississippi and South Sea companies as
well as 20th century stock exchange booms and busts using
Kindleberger’s definition (1991). The discussion of these cases,
traditionally classified as bubbles, will make clear that
characterizing them as “excessive speculation” obfuscates the
distinction between bubbles and other financial market
phenomena.
Repeated, cyclical upsurge-downfall scenarios, however, bear

several common traits that differentiate bubbles and shed more
light on their economic and behavioral background. Leverage, the
comovement of various “trust-based” speculative assets,
government intervention, media, and faith in the “new economy”
all play a role in the formation and persistence of bubbles and in
the extent of post-collapse macroeconomic repercussions. Studying
the latter is indispensable because without such repercussions the
concept of bubbles would become empty. Traditional analyses
focus on negative macroeconomic impact, but in the discussion
of our examples we will also take note of the positive impact of
stock exchange bubbles, being inseparable from negative effects.
Within the framework of literary economics the general
characteristics of stock exchange bubbles can be listed with the
capacity to foretell that stock prices are going to reach a ceiling.
When this happens a crash ensues, being the most important quality
of a bubble.
The second chapter of this paper will examine speculative

decisions – the role of so-called noise traders. These are market
VIII
actors who base their investment decisions not on relevant

information but noises (irrelevant information). Still, their
behavior cannot be considered fully irrational. Following this, we
will analyze overconfidence and the resulting “illusion of
knowledge”. Besides leverage, this “illusion” may be the chief
reason why “noise trading” becomes periodically dominant on
stock exchanges where, as a result, prices may collapse without
any relevant new public information. We believe the analysis will
make good use of the conceptual framework, whereby noise
trading will become periodically dominant on stock markets
reflected by a change in the behavior of market actors. Among
others, the unreasonable comovement of shares might be a signal
of this, the degree and change over time of which can be monitored
using an indicator. At the end of the chapter we will reformulate
previous criteria, providing a valid and practicable definition for
stock market bubbles as well as their characteristics.
The third part of the book explores the Hungarian stock market
in light of the new bubble criteria and characteristics. Special
attention will be paid to large-scale booms and busts which
occurred in the Budapest Stock Exchange (BSE) at the end of the
1990s, analyzing the underlying reasons. Finally, we aim to provide
an answer to the question whether, assessing the last 8 years of the
Hungarian stock exchange, can we talk about a stock market
bubble?
Different Aspects of Bubbles 1

Different Aspects of Bubbles
“Economics is a set of tools, from which the economist has to select

the appropriate tool or model for a given problem”.
Charles Kindleberger (2000. p.220).
6 he phenomenon of stock market bubbles is explained in a

number of ways by economists of different approaches,
depending on whether they explore it using the tools of
mathematics or literary schools of economics. Although the theses
and conclusions of these two approaches do overlap in many aspects,
it is still notable that there is hardly any substantial cross-referencing
between representatives of the two sides. While the term asset price
bubble is used in the mathematical context, analyses of literary
economics prefer the unqualified form of bubble. For this reason,
the financial encyclopedia The New Palgrave Dictionary of Money
and Finance provides two sharply different, unrelated definitions.
Asset price bubble refers to the discrepancy between the real value
and actual listing of a share (Gilles and LeRoy 1992), while bubble
in the literary economist’s interpretation is a broader economic
phenomenon where the continuous rise of share prices is fueled by
2 ANATOMY OF STOCK MARKET BUBBLES
the investors’ expectations of further increase. This interpretation

also assumes macroeconomic consequences (Kindleberger 1991).
Although used in different context, the same phenomenon is found
in the background of both definitions, namely the unlinking of
share prices and economic fundamentals.
After looking at the two-fold definition of bubbles, we will

examine the conditions for their formation and collapse, also
observing the degree to which previous research is applicable in an
analysis of stock market events.
1.1 Asset Price Bubbles in a Mathematical Approach

If interpreted for the price of a single share, a bubble is called an
asset price bubble. The mathematical definition of an asset price
bubble uses the fair price of a financial asset as its starting point.
This theoretical price is the present value of the future cash flow of
the asset. The equilibrium condition of asset pricing models is
pt = Et (dt+1 + pt+1 )/(1+ r) (1.1)
where dt is dividend, pt is the price of the asset at t time, and

Et(·) will provide the expected value of the expression based on the
information available at t time. If the interest rate (r) is considered
constant throughout the whole period, then share price at t time
(pt) in a general form can be given as follows:
pt = ∑ j =1 Et (dt + j )/(1 + r ) j + bt ,
∞
(1.2)
The first factor on the right side of this formula, which is the
discounted present value of dividends, will provide the fundamental
value of the share ( pt* ). The remainder (bt) is a deterministic or
stochastic component satisfying the condition bt = Et(bt+1) /(1 + r),

which is the asset price bubble itself.
Regarding stock investments, Keynes’ General Theory makes a

distinction between an enterprise that has the objective of predicting
the return of a capital asset during its full lifetime and speculation
that is an activity to predict the psychology of the market (Keynes
1936). The former definition therefore refers to the reservation price
of enterprising investors with a buy and hold strategy. This is the
fundamental price of the share
pt = Pt* + bt , (1.3)
and if pt ≠ pt* then in the mathematical sense an asset price

bubble is formed, where pt* is the fundamental value.
This definition, however, needs some refinement. Formula 1.3

is not valid for any and all parameters. It has to be taken into
account, for example, that the price of a financial asset cannot be
negative. Since d t ≥ 0 and r ≤ 0 , it is obvious that pt* will not be
negative, so we have the option of either not to interpret negative
bubble (bt ≥ 0) or to stipulate that bt ≤ pt* . Similarly to the majority
of models, at this point we exclude the existence of negative bubbles.
We will get back to the theoretical possibility of such bubbles in
Chapter 1.3.
1.1.1 Rational versus Speculative Bubbles

Econometric models assuming rational behavior on the part of
market actors divide bubbles into two sets. We are talking of a
rational bubble when the market price of an asset is higher than its
fundamental value, but rational expectations of market players may
justify such a price. In this event, the actual price uses fundamental
value as an anchor (Sornette 2003) staying attached to it. On the
other hand, it is a matter of speculative bubble if market price and

fundamental value diverge “too much” and there can realistically
be no dividend income that may support current market price
(Gilles and LeRoy 1992; Brunnermeier 2001). This detachment
usually means that the bt bubble component in Formula 1.2 grows
faster than the interest rate. It is a question of vital importance
whether we can clearly find a separating condition for these two
types of bubbles i.e., the critical degree of divergence from
fundamental value.
Asset price bubbles form in the presence of rational expectations

only when shares do not have a definite maturity, since in this case
rational actors could calculate the current price of the share from
its maturity price. Given such circumstances, the actual quotation
of these financial assets would always be equal to their fundamental
value. The first condition, therefore, is for the financial asset to be
without maturity, which does apply to shares. If share price diverges
from fundamental value only temporarily, then, using Blanchard’s
(1979) term, a rational bubble forms. Blanchard’s (1979) model
assumes rational expectations and a lack of arbitrage. In such a
setup, market actors recognize the divergence from fundamental
value but they expect further price increase. As in the market there
never is a guarantee for the eventual offset of prices, the bubble
keeps swelling until, within a definite time frame, it implodes and
the process starts again. The model of Froot and Obstfeld (1989)
demonstrates using the US stock market data that price fluctuations
can mostly be traced back to internal factors (dividend volatility)
but there is no method for predicting the discrepancy. That article
also underlined the theoretical problem that, in the presence of
infinite term arbitrage and completely rational actors, bubbles
cannot form.
Rational bubble models determine an infinite number of

equilibrium points and fundamental equilibrium is only one of
them. It is impossible, however, to identify the type of current
market equilibrium (Weller 1992). In fact, it is impossible to
determine what the conditions for the formation and collapse of
rational bubbles are. Tirole’s (1992) study showed that, assuming
rational expectations and a finite number of participants, in a market
equilibrium a bubble cannot exist. If any one market actor recognizes
that current share price differs from fundamental value, the only
complete rational strategy is the immediate selling of the share.
However, if there is a finite number of market players, it is not
possible for everyone to leave the market at the given price level, so
either the immediate drop in the share price will eliminate the
bubble or there will inevitably be people “stuck” with their
investments. The latter happens when a market actor will not
liquidate his investment at the current price preferring to keep the
asset infinitely. However, the finite number of participants in the
market also implies a limit for market demand. If this knowledge is
shared by all players, a bubble will not form, or will quickly
dissipate. In the context of a complete market and rational stock
market bubbles cannot exist.
The weakness of the approach presented above is the implicit

assumption of complete information supply. This aspect (the issue
of information) needs further examination because the cases
discussed above have not supported the possibility of bubble
formation and subsistence.
Up to this point in our discussion, the existence of a bubble

was common knowledge for market players, which also implied
that they could access the same amount of (symmetrical)
information. If we lift this restriction, market actors still remain
rational but they base their decisions on differing (asymmetrical)

information. The model of Allen, Morris and Postlewaite (1993)
prefers to distinguish between expected and strong bubbles instead
of rational and speculative ones on the basis of the gap between
current price and fundamental value. This means no substantive
difference in terms of classification.
An expected bubble is to be construed as a scenario where actors

recognize that the price of a share is higher than its fundamental
value but they expect further increase. Brunnermeier (2001)
emphasizes the difficulty of defining fundamental value in this case;
however, compared to the previous distinction here we can identify
additional separating conditions. The first condition for the formation
of an expected bubble is that the market is not in a Pareto-optimum,
i.e., further transactions can improve asset allocation. As the second
condition, Allen, Morris and Postlewaite (1993) stipulate that market
actors must have a short sale constraint in the future. These two
conditions are both necessary and sufficient for the formation of an
expected bubble (Tirol 1992; Brunnermeier 2001).
We are talking of a strong bubble when there can be no future

dividends justifying the current price. This condition, however, is
difficult to translate into practice since corporate earnings are
usually not possible to forecast precisely. Assuming the possibility
of estimating future dividends, a strong bubble might form when
actors sense the excessive rise in prices but they believe this is not
common knowledge. The necessary condition following from this
is that market actors must have private information, and this applies
when we can at least state that they do not know all transactions.
Brunnermeier (2001) also mentions there must be at least three
players in the market.
If, however, a market actor does not possess information in

sufficient quantity or quality, a herding effect may emerge even in
the presence of rational expectations, resulting in the detachment
of share price from fundamental value. If market actors base their
decisions on the earlier decisions of others instead of their own
private information, it may set off an informational cascade. In
Banerjee’s (1992) model, if the first two participants make the
same decision, the optimal decision for the other actors is to imitate
their behavior. If the original decision is based on false signals
received by the players, the crowd starts off in the wrong direction
(Brunnermeier 2001), while the opposite is true in case of correct
signals. In the model of Cont and Bouchaud (1998), where players
make their decisions simultaneously, herding works the same. Their
study relates this effect to the so-called fat-tail distribution
characterizing stock market returns. Herding effect applies when
market actors do not reckon with the fact that the decisions of
others are informational externalities for them. Complete rationality
on part of the participants would still arrest the formation of a
bubble. This would also require the assumption that market actors
are not imitating the decisions of other market actors, not even
those with high professional reputation. In other words, this
requirement means that there are no stock market gurus whose
decisions influence market behavior regardless of their performance.
As long as an investor will make his decision not entirely
independently from other investors, herding will inevitably occur.
That way individual decision remains rational but the sum of them
will be irrational behavior (see also Le Bon 1982). The degree to
which one will imitate other investors is also a function of the
confidence in the person’s own capacities and information as well
as those of others. To examine this, however, it is essential to analyze
the behavior of market actors, since herding cannot be deduced
from rational behavioral patterns. Chapter 1.1.3 will discuss the
influence of these factors.
Assuming rational expectations, therefore, share price will

fluctuate around the fundamental value, which means asset price
bubbles will form. There is, however, a practical difficulty regarding
the applicability of the concepts of expected and strong bubbles.
The above models indicate that both expected and strong bubbles
(which are not really distinguishable) are created on a recurring
basis. It is a rare occasion, and an occasion not even possible to
spot, when the price of a share equals its fundamental value. As we
have seen, market price may well be higher than this value with
perfectly acceptable reason. This means that, based on the formula
for share price (1.3), the market price of a share always contains a
bubble component, which, in turn, may lead us to the conclusion
that the analysis of asset price bubbles is irrelevant, all previous
prices being justifiable in hindsight (Garber 1990, 1992, 2000;
Flood and Hodrick 1990).
Until now we have focused on the internal causes of bubble

formation, realizing that internal instability is inevitable when
market actors possess differing information. Still, the existence of
rational bubbles cannot be deduced from this without
contradictions. External uncertainty factors, collectively called
sunspots, however, do have the potential to provoke high fluctuations
in fundamental value. Sunspots are external events with no effect
on an investor’s preferences, informational supply or capacities, but
as signals being received by all market actors, they will affect
individual convictions about the behavior of fellow players
(Brunnermeier 2001). In the model of Allen, Moris and Postlewaite
(1993) external uncertainty caused by sunspots will push the
economy from one equilibrium position to the other. Similarly to
other approaches, however, this model has the drawback that, while
providing a valid explanation for bubbles using the assumption of
rational expectations, accepting the existence of sunspots weakens
the economic practicability of models integrating them, because

sunspots do not lend themselves to detailed analysis.
Studies assuming rational behavior teach us two important

lessons. First, short sale constraint is a prerequisite for the formation
of rational bubbles. Stock owners will someday have to sell their
shares, so they will inevitably re-enter the market as sellers. This
infers that nobody can, and will, hold his shares indefinitely,
therefore, no liquidity constraint is created, triggering an infinite
rise in prices. This raises the question whether in our analyses we
can reasonably use fundamental value, which is, after all, a concept
assuming infinite possession of the share (what we have called a
“buy and hold” strategy). To address this concern, the next section
explores the issue of fundamental value.
1.1.2 The Problem of Determining Fundamental Value

As it was shown, the fundamental or internal value of a share, which
is the point of reference for identifying bubbles, is determined by
future dividends. However, this starting point has a logic flaw since
the definition of fundamental value provided in formula (1.1), based
on a “buy and hold” strategy, fails to account for trading motives
other than speculation (i.e., risk-taking in hope of extra return)
after buying an asset. In complete markets, where no informational
problem exists, every player can assign a payoff to possible future
positions and make a hedge, thus ensuring Pareto-optimal allocation.
If the market is incomplete, it is possible to conceive a position to
which individual actors cannot assign a payoff. In such cases, either
information is asymmetrical or there is a lack of priors, so only
trading the exchange of stocks will ensure a Pareto-optimality. Here
trading is not speculative and on the basis of the definition for
fundamental value an asset price bubble inevitably forms.
Brunnermeier (2001. p.49) uses the following example to

illustrate why liquidity and free trading in shares may elevate share
price (see Figure 1.1). Two types of actors in an economy (X and Y)
t=0 and t=1 each owns a risky asset, which pays dT dividend at the
end of the second period (t=2=T) in an economy being in ω state.
We assume ω ∈ Ω = {ω1 ,ω 2 ) , i.e., the whole set of states consists of
two states. The probability of these two states is equal and the
actual state becomes known before the t=1 trading period. The X
type actor values dT (ω1 ) dividend equal to 1 unit in state ω1 and
0 unit in state ω 2 , with opposite preference for Y type actors. If we
disregard the time value of money, it is deducible that actors will
value the dividend at 0,5 unit in period (t=0) if they hold on to
their shares until dividend payments (i.e., the second period).
Figure 1.1: Stock Value if There are no Trading Opportunities
Actors t=0 p t=1 t=2 Ω

½ 1 1 ha ω1
X 0,5
½ 0 0 ha
ω2
½ 0 0 ha ω1
Y 0,5
½ 1 1 ha ω2
Note: solid arrows indicate backwardation, and probability of states p = ½.
If, however, they have a theoretical opportunity to trade at the

end of the first period (t=1) and, according to the above condition,
the actual state of the economy becomes known prior to this, then
the value of the dividend may be 1 unit in the zero period, because
trading allows the type of actor having a higher evaluation to buy

the share (see Figure 1.2). If state ω1 turns out to be the fact, X
type actors will pay 1 unit for the shares of Y type actors (dotted
arrow in Figure 1.2). Of course, trading might occur at a different
price, but for the sake of simplicity we assume that the buyer offers
the highest price. If the actual state of the economy is ω2 , trading
will occur in the opposite direction. Different a priori opinions will
make it possible for the actors to value the dividend at 1 unit from
the beginning, since they have different evaluations of the dividend
in the same state of the economy. This special case, therefore,
provides for an immediate bubble depending on the share structure
because not all shares with optimal payoff for market actors are
available in the market.
Figure 1.2: Stock Value if There are Trading Opportunities
Actors t=0 p t=1 t=2 Ω

½ 1 1 ha ω1
X 1
½ 1 0 ha
ω2
½ 1 0 ha ω1
Y 1
½ 1 1 ha ω2
Note: solid arrows indicate backwardation, dotted arrows show trading with
shares, and probability of states p = ½.
Brunnermeier’s (2001) explanation: through this mechanism

investors are completing the market by achieving a payoff not
available to them without trading. In this way the trading space
expands and the new equilibrium point will be Pareto-optimal.

This example demonstrates that trading may have a motive other
than speculation. In practice, the two motives are indistinguishable
from each other since in an incomplete market both can be a source
of trading, reflected in market liquidity.
One of the main functions of the secondary market is providing

liquidity, which invariably elevates share price. Let us assume a
public company is about to raise capital and has two alternatives.
One is to target a narrow investor group with the new shares, which
will be issued privately and only later listed in the stock exchange.
The other is to offer common shares for a wide investor group,
listing them immediately in the stock exchange. In these two cases,
the issue price will be different due to different liquidity. If we are
thinking in terms of a “buy and hold” strategy then fundamental
value is the same in both cases and liquidity affects price. (We will
get back to the question of how risk affects fundamental value.)
Silber (1992) classifies the scientific answers given to the question

of what return market liquidity offers. Liquidity characterizes
trading but it does not only refer to sheer volume. Liquidity means
more or less that the given asset can be sold quickly and simply
while market equilibrium price changes only slightly. A liquid stock
market is deep because sell and buy offers are close (bid-ask spread
is small) and even high volume sales do not influence price
significantly. In studies, the most widely accepted method is to
analyze the impact of two internal components: bid-ask spread
and the development of the price carrying a liquidity premium
(Schwartz 1992). Several studies about these two components lead
to similar conclusions: capital cost for a company is undoubtedly
reduced if the stock has a liquid secondary market. It has been
proved that publicly offered stock without any limit on trading is
worth more than illiquid stock, which, on average, has a price 30%
smaller than listed shares with similar business risk and shareholder
rights (Silber, 1992).
The argument above shows the economic return provided by

stock markets through improving capital allocation. We could also
see that this increases the price of the financial asset. Previously,
judging from the Keynesian definition of enterprise, the objective
of possessing shares and the fair value attached to this possession
could seemingly be defined in terms of future dividend income.
Speculation was set apart as a necessary evil causing instability in
the market. Keynes even suggested in his General Theory, referring
to the consequences of the 1929 stock market crash, that the
prohibition of buying investments might have cured some actual
problems, because it would have forced investors to focus on long-
term prospects (Keynes 1936). But the General Theory also argues
that the illiquidity of investments may restrain new investments as
well, and this would evidently result in a price depressing effect.
This means speculation (secondary market-based trading) has
undisputable benefits besides the damage caused by instability.
The two effects are both inherent to stock markets, and this is the
only way to broadly share and spread risk.
Tirole (1992) defined four benefits of speculation in accordance

with the argument laid out in this chapter. First, we can mention
the insurance opportunity emphasized by Keynes (1936) and Hicks
(1946), which is utilized by a company when it offers stock in the
primary market. This is practically a hedge transaction where the
enterprise sells a part of its future profit to new shareholders while
spreading (sharing) its risks. Primary markets, therefore, help the
growth of companies because the issuer parts with a portion of its
expected profit in exchange for reducing business risk. The second
share price-enhancing benefit of speculation is the above discussed

secondary market liquidity offering shareholders the opportunity
to sell their risks to other investors (speculators), who willingly
take it in hope of future profits. The third benefit is that the
continuous market presence of speculators provides a way for
investors to sell their shares if they value the attached income to be
smaller than current price – thereby providing a put option. If
there were no investors in the market with speculative intentions,
trading would most probably be minimal (Black 1986) or non-
existing. The fourth benefit comes from the speculation (risk-seeking)
of financial intermediaries or portfolio managers, which will have a
positive return for other market players. If only one share is traded,
then, in theory, this can be interpreted as reduced risk for market
players resulting in the increase of price and fundamental value.
Weil (1990) deduced this way that price can lower fundamental
value, which also implies, however, the mutual dependence of
dividend and price. Assuming investor rationality, we can again
conclude in this case that bubbles do not exist.
Speculative stock market is often compared to a financial Ponzi-

scheme. This refers to an investment scheme whose workings make
it impossible for all players to reap monetary rewards but every
player is sure that he still entered the game in time. This
characterized the pseudo-enterprise originally started by Charles
Ponzi in the 1920s, which aimed to realize a profit on the “arbitrage”
resulting from the difference between exchange rates and between
the fixed European and the US prices of international postal reply
coupons. The profit promised by Ponzi, however, could not have
been achieved even if the enterprise sells all the reply coupons in
the world (Shiller 2000. pp.65-66). In the case of a listed company,
however, rarely can an obvious profitability limit be given, and this
is even more so in the case of a new industry promising cost
efficiency and tangible benefits in production, services and sales.

In case of borrowing, a Ponzi-scheme can be recognized if the net
cash flow of the investment is not enough even for interest payment
(Tarafás 2001) but the game will not collapse as long as new loans
(new investors) provide funds for immediate cash flow needs. This
is the case of blank credit. When turning to the stock exchange, we
cannot speak of Ponzi-financing because share price increase is
equivalent of dividend payment for shareholders. If the company
makes investments with a positive net present value, share price
will theoretically rise, and only in hindsight, in light of the failure
of the investment we can be absolutely sure that the fundamental
value of the share offers no collateral for capital re-payment in the
form of dividends. This means fundamental value cannot be
determined in advance. In other words, our analysis has no use for
the concept of fundamental value.
This chapter reviewed the theoretical problem that dubious

definitions for fundamental value and asset price bubble strongly
reduce the practical relevancy of models. This is the biggest obstacle
in the way of empirical tests for speculative bubbles (Flood and
Hodrick 1990), which means the existence of asset price bubbles in
the stock market cannot be verified within the framework of
mathematical economics. In the following chapters we will broaden
the framework of the analysis to include not only mathematical tools
but also methods of a psychological approach. The question of stock
pricing can also be analyzed by taking into account the specifics of
investor behavior and its effects, confirmed by empirical studies.
1.1.3 Behavioral Finance on Bubbles

So far we have assumed rational behavior from market players and
took it for granted that prices reflect all public and private
information. This can be re-phrased as, based on available
information, “shares are priced correctly”. The basis for the Efficient
Market Hypothesis (EMH) is that the price of a financial asset is
affected by unpredictable future information. Therefore, if prices
reflect available information then they will also be unpredictable,
following random walk. Barberis and Thaler (2002. p.4) illustrate
the two approaches as:
“prices are right” => “there is no free lunch”
“there is no free lunch” ≠> “prices are right”.
In an efficient capital market the two statements follow from each

other, but in other cases the validation of the latter does not necessarily
mean the validity of the first statement. This means the fact that
there is no investment strategy ensuring a return beyond what is the
compensation for increased risk does not prove correct pricing.
In the past two decades several studies pointed out the theoretical
weaknesses of EMH. The paradox of Grossman and Stiglitz (1980)
states that if a capital market is efficient then gathering information
is in nobody’s interest. Therefore, it is necessary for information
gathering to be costly. Shiller (1981), and Grossman and Shiller
(1981) claimed, on the basis of empirical data, that dividend
volatility alone does not explain the fluctuation of share prices,
which means share price fluctuation is not closely related to
fundamental value. This contradicts the findings of Froot and
Obstfeld (1989), whereby price fluctuation can rationally be
supported by the fluctuation of dividends. Shiller (2000) says
co-movement of dividend and price does not mean a rational
behavior because both might reflect the irrational behavior of the
market. Shiller (1991) investigates an alternative explanation for
the problem which he could not confirm eventually whether the
volatility of expected returns may account for excessive price
volatility. Analyzing the future prices of orange juice, Roll (1994)

concluded that prices are not moved by news, which is not consistent
with the assumption of rational investor behavior (i.e., appropriately
reacting to information).
The validity of EMH has been the focus of another field of

research, known today as Behavioral Finances (BF), which is one of
the most important branches of finance in the last one and a half
decades. Instead of assuming rational expectations, BF explains share
price fluctuations with the regularities and psychological patterns
of investor behavior. A large number of empirical research found
anomalies weakening the practical validity of EMH.
BF research concentrates, therefore, on why incoming

information is not integrated correctly into prices and what the
reasons behind incorrect pricing are. If two shares are fundamentally
identical and their market prices differ, the strategy of a rational
investor would be the following: open a short position for the
higher-priced and a long position for the lower-priced share. If
market players are completely rational, this opportunity for arbitrage
disappears immediately following from the identical decision of
players. The arbitrage based on the relation between the
fundamental values of two shares is similar. In the stock market,
however, no perfect arbitrage can be done in either case. Arbitrage
over time is always costly and risky (Barberis and Thaler 2002), as
there is no guarantee for the equalization of the two prices within a
relatively short term. The reason behind this is that non-rational
investor behavior might play a decisive role in price development
during a longer period.
Non-rational investors are in fact following typical behavioral

patterns, so their decisions are not entirely wrong (i.e., irrational),
rather quasi-rational (Thaler 2000), because the sources of guiding

patterns can be identified. Our decisions are usually helped by
certain “rules of thumb” in choosing the most effective (i.e., quickest
and most precise) option. The literature of BF does not yet offer a
uniform map of psychological factors because there are overlaps.
The most important characteristics, however, are excessive confidence
and optimism, heuristics, anchoring and conservativism, framing
and cognitive dissonance. Based on comprehensive studies on BF
(Barber, Odean and Zhu 2003; Barberis and Thaler 2001; Shleifer
2000) as well as Komáromi (2002a, 2003c), these patterns and
their practical effects are summarized in Table 1.1.
BF does not focus directly the emergence of bubbles because it

examines prices in terms of how much they reflect available
information. It traces the digression from fundamental values,
which, as seen earlier, is an asset price bubble from the mathematical
point of view, back to typical investor behavior. Incorrect market
Table 1.1: Main Behavioral Patterns in Investors Decisions

Behavioral patterns Effects
Optimism overconfidence Investors are more active, volume increases
in abilities/information Overreaction to news
Representativeness Overreaction to rare events
Wrong and overreaction because
similarity
Conservatism Anchoring Underreaction because investors do
not update their beliefs
Framing Risk attitude, direction of decision
depend on the interpretation of the
news, situation
Cognitive dissonance Investors tend to follow trends:
herding.
Source: Odean (1999), Shleifer (2000), and Shiller (2000).
pricing can be recognized empirically and confirmed from the

psychological side. For example, the model of Barberis, Shleifer
and Vishny (1998) illustrates the phenomenon that investors
underreact information in the short run and overreact in the long
term. This can be traced back to two factors included in Table 1.1.
At first, the new information regarding the given share does not
change previous conceptions enough as investors are conservative,
“anchoring” their decisions to their former point of view. This leads
to an underreaction to the news in the short term, because fresh
news are weighted less in investor decisions. Later, a series of good
or bad news might cause investor opinion to follow a characteristic
heuristics, resulting in an extrapolation to the future and a
long-term overreaction to the news.
In the model of Daniel, Hirshleifer and Subrahmanyam (1998),

trust in private information grows if it coincides with public
information, and weakens if they contradict. The short and long
term fluctuation of market prices can be deduced from this. In this
model, the change of trust in information and personal capacities
can be explained by cognitive dissonance. Every private decision
invokes doubt on the part of the actor about whether he has made
the right decision; therefore, his attitude will not be consistent
with his behavior. This doubt, or inconsistence, is cognitive
dissonance itself, the reduction or elimination of which is a natural
human reaction. Investors try to justify their buying or selling
decisions in hindsight. According to the above model, if a later
published information confirms earlier private information of an
investor, his self-esteem will grow, his cognitive dissonance decrease.
The two basic models of behavioral finances discussed above,

however, do not explain the general phenomenon that, based on
50-70 years of data from US stock markets, shares follow previous
trends for a couple of month (momentum effect) then they will
revert to the average in subsequent years (reversion effect). To sum

up what we have examined so far: BF describes reality more precisely
than EHM because it takes behavioral specifics into account. Now
we see that the formation of a bubble strongly depends on how
investors perceive and to what degree use new information, and
what psychological patterns are observed during their decisions.
Still, the definition of fundamental value as a benchmark is a weak
point of these models. With the help of assuming limited arbitrage,
we obtain an explanation for why the price of two fundamentally
identical financial assets might diverge, like the share price of a
closed-end fund and the value of the stock portfolio managed by
the fund. However, with the examination of the interrelation of
financial assets, BF partly circumvents the issue of fundamental
value, not providing the necessary and sufficient conditions for the
formation of a bubble. BF still proves that psychological
characteristics can partially explain share price development, but
further thought needs to be devoted to what role investor behavior
may play in the formation of bubbles. We revert to this in
Chapter 2 where we discuss two phenomena: overconfidence and
the illusion of knowledge.
Unfortunately, BF models do not offer an answer to the question

as to whether we can differentiate between a large aberration from
fundamental value and a smaller bubble. Furthermore, BF literature
explores the development of most of the psychological characteristics
only in general terms, so we cannot tell whether investor behavior
changes from one period to the other. We are not able to compare
behavioral characteristics of market actors at the beginning of a
boom with those during a market crash. Psychological patterns
laid out by BF, therefore, cannot explain the formation or collapse
of bubbles directly.
1.1.4 The Predictability of Crash

The introduction of the mathematical approach to bubbles would
not be complete without answering the question whether the
bursting of asset price bubbles is predictable. The models shown
above sought to find an answer to how a bubble forms, in what
circumstances, and if it can be explained by rational or irrational
behavior. According to EMH, share price fluctuation is
unpredictable, and although empirical research related to BF did
find some past information which might make prices more or less
predictable, this does not modify the message of EMH significantly
(Komáromi 2002a). If share price diverges too much from
fundamental value, all of the above models assume that it will return
to it. However, this cannot be proved in practice, in fact, not even
whether a change of fundamental value justifies a previous price
change. Approaching the problem from a purely mathematical-
statistical point of view, it can be stated that the distribution of
stock returns has fat tails regardless of whether we examine the
possible evolution of theoretical value or actual stock market data.
This brings about the natural instability of market prices. A stock
market is more unstable compared to other complex systems because
decisions here are made by individuals whose personal considerations
multiply the number of factors to be reckoned with.
About a decade ago, the first articles written by physicists

pointing to regularities in the chaos of stock market decision-making
were published. It has been proved that the distribution of stock
returns within a monthly (or shorter) period differs from normal
distribution and their deviation cannot be given (they have stable
Lévy-distribution). Farmer (1999) also argued that for longer
periods, normal distribution is a good approximation, where the
second momentum can be calculated. Moreover, it can be shown
that, examining a large amount of data, these returns have

asymptotical power law characteristics. Sornette’s (2003)
comprehensive study tried to apply techniques borrowed from the
science of physics to capital markets, emphasizing the imperfections
of their predictive power. In his research, he determined the critical
time when prices begin to fall by using linear and non-linear
equations. Fifty percent of the case studies listed in his work contain
false predictions, which may even be considered significant
depending on one’s approach, but definitely not economically
relevant from our point of view.
Farmer (1999) stresses that such research does not yet have a
sound economic background but it undoubtedly points out some
regularities. For the time being, however, the use of past prices to
predict future ones provides contradictory and irrelevant results.
For instance, Lo, Mamaysky and Wang (2000) found some patterns
of the technical analysis in US financial data but without any
significant explanatory power. All we get here is technical analysis
with a theoretical coating, instead of economic or behavioral
regularities. One dynamically developing branch of the physical
approach, the science of networks, cannot be applied to stock market
analysis efficiently, but a subdomain of it, research on not-random
networks shows promise in terms of mathematical modeling of
collective decision-making (e.g., Barabási 2002).
1.1.5 Conclusions
The previous subchapters discussed asset price bubbles from the
theoretical and practical point of view but we were left facing several
problems. Exact model-making is both the advantage and
disadvantage of the mathematical approach because although this
presents several market characteristics, limiting conditions make it
valid only in some specific context. This is the conclusion of a study
by Flood and Hodrick (1990): asset price bubble models have an

interpretation problem making the verification or discarding of an
asset price bubble difficult.
In spite of such contradictions the examination of asset price

bubbles offers several insights. First is the demand-side limit on
liquidity, which is trivial, still it is the most important condition
for the formation of an asset price bubble. Tirole (1992) compares
shares to “hot potato”, passed from one player to the other, with
the last one unable to get rid of them. This means there will always
be players stuck in their investments, which is equivalent to what
Keynes (1936) expressed that there is no such a thing as liquidity
of investment for all the market players. However, even if the investor
cannot sell his share, he receives dividends so it is not a Ponzi-
scheme in the classical sense, only if there is no fund for dividends.
Or to put it more precisely, the value of corporate assets and the
present value of the future return on investments do not justify
shareholder expectations in hindsight. There may be two reasons
for this, the first being the information problem.
Market actors do not make their decisions based on complete

information. So there is a high probability of a strictly interpreted
bubble in the case of any given share. In light of what has been said
so far, a rational (expected) bubble is not distinguishable from an
irrational (strong) one. In the latter, case we could put forward the
criterion that no future dividend income can justify the current
price. The only problem here is that the correctness of predictions
regarding corporate profits is only confirmed ex post, as it is hard to
define a clear limit in advance. Such an exceptional case is the Ponzi-
scheme but this is uncharacteristic of stock market investments so
it is not applicable for the explanation of bubble formation and
collapse. Section 1.3 uses the historic point of view of literary
economics to expand on this issue because it is easier to explore the

problem in specific cases. Another issue with regard to “unrealistic”
return expectations is the principal-agent problem, i.e., the role of
management in shaping expectations. This important aspect will
also be the topic of Section 1.3.
Investors will invariably sell their shares in the future (short sale
constraint) and so they cannot be divided into entrepreneurs and
speculators in the Keynesian sense. These two groups of shareholders
are only separated by differing investment periods. During the
examination of bubbles, therefore, the question is, why do they act
as sellers in the market in a given moment? The second chapter
explores the factors which imply a ceiling for further price rise and
the short term inevitability of a crash.
The second consequence was highlighted by the literature of BF:

recognizing and validating the fact that there is no perfect arbitrage
in the market. This inevitably implies that the price of a particular
financial asset will rarely match its fundamental value. Fundamental
value cannot be used as a benchmark because price fluctuation is
mainly a function of investor behavior. Changes in the latter will be
examined in Chapter 2 with regard to overconfidence.
The next section presents the phenomenon of asset price bubbles

from a different angle, through the prism of empirical economics.
1.2 Asset Price Bubbles in Laboratory

Empirical economics enriched the science of economics with a wholly
new approach. It directly examines the factors influencing investor
behavior in laboratory experiments with specific, pre-defined
conditions. One arm of this approach is the creation of artificial
markets and conducting behavioral simulations in laboratory. The
following will summarize the results of such research.
During an experiment, separated participants will buy or sell

shares in a given number (usually 15-20) of rounds, determining
the price observable in the monitors. The subjects do know the
timing and value of dividend payments, and they are aware that
prices are determined only by investor decisions. It is common
during these experiments to grant the profits to participants,
ensuring profit-making motivation. These experiments are repeated
at least 20 times to ensure statistical validity or control experiments
are conducted to complement the laboratory study.
The article by Smith, Suchanek and Williams (1988) presented

the classical standard for the following capital market experiments,
where the share paid a given amount of dividend after a pre-defined
number of rounds. Subjects, therefore, knew the fundamental value
of the asset, still prices at first usually went up during trading then
sank, and although they never detached themselves completely from
fundamental value (preserving the validity of the condition of rational
expectations), in the mathematical sense an asset price bubble was
formed. The experiment also showed that participants could forecast
next-round price with only a small degree of precision. During price
increases, they usually underestimated future prices and vice versa.
Trading volume was significantly influenced by price fluctuation:
the activity of participants was higher in times of price increase, and
smaller when prices fell. The most important result presented by the
authors was the realization that the size of an asset price bubble was
dependant on the experience of participants. If they have taken part
in several experiments previously, then divergence from fundamental
value was significantly smaller, even if it did not disappear entirely.
They believe the reason behind the formation of a bubble is the
diversity of experiences and capacities of participants, in spite of their
common knowledge about dividends.
Stanley (1997) explored further the possibility of irrational

bubble formation during laboratory experiments. His experiments
were different from earlier ones in as much as he did not specify
the number of rounds. This increased uncertainty causing a higher
divergence of prices from fundamental value and a formation of
speculative bubbles in the sense used in previous chapters. He
observed a divergence from fundamental value even in those rounds
where any further dividend payment was out of question.
Lei, Noussair and Plott (2001) explored the option whether an

asset price bubble will form in the absence of speculation. Their
experiments also created asset price bubbles which they explained
with the constraints on the activities of participants: they could
buy or sell shares but could not do any other activity. They had the
option not to trade but it is difficult to withdraw from active
participation in an experiment (i.e., trading). This touched upon a
weak point of empirical economics: experimental circumstances
influenced the behavior of participants per se, causing them to act
differently from how they would react to a real-life investment
opportunity. These experiments also offered an answer to another
question: if dividend information is common knowledge, why does
it not stop bubbles from forming? In an experiment, there are always
inexperienced participants who need time to understand the formal
and informal “rules”. Similarly to the experiment by Smith,
Suchanek and Williams (1988), a bubble does not disappear even
with experienced participants, for which they offer the explanation
that the rationality of participants is not common knowledge. The
hypothesis by Lei, Noussair and Plott (2001) was that participants
at first do not assume the others to be rational, which leads to the
formation of a bubble. As soon as the rationality of all players is
evident, prices begin to fall and the asset price bubble disappears.
The previous studies concentrated on whether asset price bubbles

form and if they do, are they caused by rational or irrational behavior.
The validity of these experiments is undermined by the fact that
during such experiments participants expect bubbles to form simply
based on the specificity of the situation, and usually rightly. This
inevitable experimental flaw can be circumvented by checking, ceteris
paribus, the change in probability of bubble formation and in
bubble size depending on the modification of one particular
condition. So far, two factors have been identified as causing a large
divergence from fundamental value: participants with limited
experience, and the lack of certainty about the number of rounds
in the experiment. Caginalp, Porter and Smith (2001) demonstrated
the effect of other factors which are summed up in Table 1.2 along
with the previous ones.
Table 1.2: Factors Determining the Extent of Price Bubbles

in Laboratory Experiments
Great price bubble Small price bubble
1. Subjects have less experience Subject have more experience
2. Number of rounds are not known Number of rounds are known in
in advance advance
3. High cash Low cash
4. Dividends are paid immediately Dividends are paid with a delay
5. No information about the transactions Information about the
transactions
Source: Smith et al. (1988), Stanley (1997), Caginalp et al. (2001).
The effect of the first three factors is stronger and better validated
from a statistical point of view. The first factor increases uncertainty
as participants are less certain about the outcome of the experiment
and they expect (speculate) that the others will not decide rationally.
The second factor also increases uncertainty because the time frame
of the experiment is not known; the participants have no

information about the number of trading rounds. The third factor
verifies the role of liquidity in bubble formation. Caginalp, Porter
and Smith (1988) recognized that if a new experiment provided
one dollar more per share for the participants, the maximum price
of the share was more than one dollar higher in most cases. The
latter two factors have a smaller explanatory power, nevertheless
they influenced the size of the bubble in many cases. If dividend
received during the rounds was credited only at the end of the
experiment, liquidity effect reduced bubble size compared to
immediate payments. It also became clear that publishing direct
information on transactions (price, volume, time) decreased the
probability of bubble formation.
1.2.1 Conclusions
The most important result of stock market experiments conducted
in laboratories is that during such an experiment trading conditions
(cash, number of rounds and participants, etc.) can be specified
and the impact of changing them can be tracked.
Although experimental circumstances may distort the

observations, they do confirm some of the factors determining
bubble formation. The first such factor is uncertainty implying a
bubble, which can be countered by providing the participants with
additional information, like publishing trading data or the number
of rounds. From among the propositions of mathematical models
the importance of liquidity is also confirmed by these experiments,
one method of which is the postponement of dividend payment.
Another statistically significant finding is that an increased initial
amount of cash leads to a relatively larger bubble.
A previously unexplored factor, the experience of participants

also influences laboratory bubble formation. The more games the
participants took part in, the smaller the bubble. It is important to

note that the involvement of relatively more experienced participants
still did not exclude bubble formation entirely, the reason for which
is not clear. It could be the distorting effect of the experiment, or
more experienced subjects have less illusion about their capacities
to beat the market. If in such cases, a bubble would disappear
within a couple of rounds, that might signal an effective market
where given the available information share price does not detach
itself from fundamental value. A positive asset price bubble does
not disappear in most cases, but trading can sometimes happen
even below fundamental values (Srivastava 1992). This contradicts
earlier assumptions about the impossibility of negative asset price
bubbles, though the reasons behind their emergence in laboratory
circumstances are still unclear, perhaps being simply the distortional
effect of the experiment. Another result not coinciding with earlier
mathematical models is the fact that a lasting bubble may form
even when the maturation date is known. This means investors
make a mistake in determining the price (i.e., during discounting),
which may imply not every decision of the participants is rational.
The fact that such lab asset price bubbles do not work off quickly
shows participants are aware of the existence of irrational decisions.
At the beginning of Chapter 1.1 a distinction was made between

bubble and asset price bubble. Mathematical models and laboratory
experiments examine the latter, which is a divergence from
fundamental value. Necessary and sufficient conditions can be
determined, but their validity is undermined by their generality.
Using the strict definition we could almost always register an asset
price bubble. The above discussed approaches do not offer an
adequate method to distinguish between speculative and rational
(expected) bubbles, either. The single-most important unanswered
question with regard to experimental capital markets is the
interpretability of their findings, the connection between “real”

asset price bubbles in real-world settings and laboratory phenomena
identified as such.
1.3 Literary Economics and Bubbles

Until this chapter, we have discussed the problems of asset price
bubbles being defined as a positive or negative difference between
actual price and fundamental value of the financial asset. The term
“bubble” is a broader concept covering an economically relevant
phenomenon affecting the whole of the capital market. While
previous mathematical approach explored how an asset price bubble
forms in the case of an individual share, literary economics sets out
to find the answer as to why stock market bubbles emerge. The
difference in the question is related to the difference in the approach,
but they both serve to explain basically the same phenomenon.
The most widely accepted definition of bubbles is given by Charles

Poor Kindleberger as: “A bubble may be defined loosely as a sharp
rise in price of an asset or a range of assets in a continuous process,
with the initial rise generating expectations of further rises and
attracting new buyers – generally speculators interested in profits
from trading in the asset rather than its use of earning capacity. The
rise is usually followed by reverse expectations and a sharp decline in
price often resulting in a financial crisis.” (Kindleberger 1991. p.20).
This definition is the essence of what Kindleberger drew from the
most important tool in literary economics: historical examples and
recognizable parallels of them. Hereinafter we are going to use this
definition to characterize stock market bubbles, the most important
parameters of which are the following:
1) Initial rise, expectations of further rises: Kindleberger (2000)
found the origins of this in an exogenous shock (displacement)
affecting the economy, modifying economic outlook in a

positive way. This can be different in different eras; either
quantitative, like the discovery of a new continent, or
qualitative, like a technical invention enhancing the
effectiveness of production.
2) New buyers: The demand for shares increases; more and more
participants take part in trading, and the activity of the
players grows.
3) Speculation: Investors do not buy with the aim of receiving
dividend income, rather price gains. Although this definition
has weak points mentioned earlier, it will be used as a starting
point in our studies, in the sense that the proportion of long-
term investors aiming to receive dividend income decreases
along with the average investment period.
4) Price decline: The collapse of prices and the whole of the
market may occur suddenly or gradually, with players leaving
the market.
5) Financial crisis: Although Kindleberger did not consider this
to be a necessary consequence, the following discussion of
historical examples will account for the positive and negative
macroeconomic impacts as well, such impacts lending an
economic weight to the phenomenon.
Traditional bubbles known by economic history will be presented

in light of the above listed five parameters in the following
sub-chapter, but the study will be complemented with common
traits recognizable in these examples, which may signal a drastic
change in investor trust and buying attitude.
1.3.1 A Classic Example: Dutch Tulip Bulbs

Probably the first and most often quoted example in modern-age
economic history is the case of Dutch tulip bulb speculation (1634-
1637), which is considered by finances literature as a classic bubble.
Even though tulip is not a financial asset, the accompanying
speculation, similar to one in a capital market, is an archetype in
economic tradition. Shiller (2000) mentions other documented
speculation from earlier centuries, for example in Far-East spices.
We believe, however, that in order to speak of speculation, a
developed market and a large number of participants is necessary,
which were out of question previously. That is why the case of
Dutch tulip bulbs may be considered as the opening chapter of a
new era in economic history as well as a permanent reference point.
A problem arises from the fact that the “yield” of a tulip (i.e., the
value of its beauty or the envy of the neighbors) does not correspond
to the dividends of a share. Still, the phenomenon referred to as
tulip mania is so much the part of the framework around economic
debate on bubbles that its analysis cannot be shunned.
17th century Netherlands was a center of international commerce

and finances of the era. The two flagships of the flourishing industry
and the fast developing economy were the Dutch East India
Company (founded 1602) conducting trade with Asia, and the
West India Company (founded 1622) conducting the American
slave trade. The ports of the country were a focal point in the trade
of many commercial goods, one of them being tulip bulbs of Turkish
origin. Market liberalization happened in 1634 (Garber 2000)
allowing tulip bulb trade for everyone. Its market was, from the
technical point of view, very similar to the already very advanced
financial markets where, in addition to prompt transactions futures,
contracts and options were also traded. The latter were, however,
not regulated in a strict legal way and fulfillment was not entirely
protected legally.
The first four parameters in Kindleberger’s bubble definition

can be interpreted in the case of Dutch tulip bulb speculation as
follows. Initial displacement was the economic impact received by
the Dutch economy at the beginning of the 17 th century, even
though the social and political environment was not nearly as
favorable with a plague epidemic and a war waged against the
Spanish. Still, economic outlook was bright, mostly because the
leading role in international commerce was already consolidated.
The arrival of new buyers was thanks to the abolishment of trade
restrictions on tulip bulbs. Starting in 1634, therefore, a new market
was born. Speculation was given a further push by the palpable
increase in demand for special, rare tulips, the new symbols of
richness, in France from the beginning of the 30s (Garber 2000).
The exclusivity was borne from the fact that, due to a virus infection,
tulips with special patterns appeared, becoming a fad among
increasingly affluent citizens.
Although the tulip bulb case is usually mentioned as an example

of irrational pricing (Kindleberger 2000; Galbraith 1994),
determining fundamental value does pose a problem. As we said
above, the internal value of a tulip, the “fundamental value” of its
beauty cannot be specified. Average prices went up 5-10 times
between 1622 and 1637, and, even though the degree of price
increase was dependant on the species, the co-movement of prices
was recognizable: the price of ordinary tulip bulbs followed the
same upwards trend. The bulb of the most famous and exclusive
tulip, the Semper Augustus, was bought for 6290 guldens in February
1637, which was a small fortune at that time in the Netherlands.
Galbraith (1994) relates one of the best-known economic anecdotes,
in which a trader invited a seaman for dinner, and when he was
looking for his tulip bulb the end of the feast, it turned out to be
eaten by the guest as a garnish to the herring. The trader’s loss was
indisputably huge, but even such argument (see Galbraith 1994)

does not prove the price of Semper Augustus was irrationally high at
the zenith of speculation. At today’s prices, one bulb was worth
about 50.000 USD (calculated with average gold price in 2003),
but those who bought at this point was only miscalculating the
tulip bulb fad and market demand. We can bring up the example
of modern-day trade of artworks, where even million-dollar
purchases are not considered irrational, simply a result of current
fashion, which can even be seen as an investment. It is not possible
to find, even as a reference point, an internal value of the tulip
bulb for Semper Augustus, either. A price of 1000 or 3000 gulden is
no more “realistic” than 6290. During a speculation, which is
actually part and parcel of every market, there will always be
participants who get “stuck” in their investments, or to put it in
another way, the market is not liquid for everybody at the same
time (Keynes 1936). Therefore, we cannot talk about irrational
pricing, only the bad luck of those buying (or not selling) at a high
price. Although the number of those who bought a tulip bulb for
its beauty was probably higher at lower prices, it was evident only
in hindsight that the scope of people only motivated by
“fundamental value” did not expand.
In the case of Dutch tulip bulbs there were a couple of factors,

from which price collapse is not directly deducible but which were
nevertheless indicators of an impending crash. In early 17th century
Netherlands many started to deal in two increasingly popular
financial transaction types: derivatives and mortgages. Although
exact numbers are not available, historical economic data implies
their rising popularity. Even though futures contracts were
forbidden by effective Dutch law, around this time the enforcement
of such rules was gradually weakening because of the interests of
market players. Of those who entered into futures or options
contracts, or borrowed money in order to trade in tulips, most

were presumably not interested in the bulb fad, rather in making a
profit from trading. There was an increasing risk that bulbs bought
from loans would appear in the market. Both Galbraith (1994)
and Bácskai (2003) identified leverage as a common denominator
of financial bubbles. This coincides with the above argument saying
the increase of such financial transactions may signal the “excessive”
rise of speculation.
The other phenomenon becoming recognizable in spite of spotty

data is that a rise in prices and speculation was not only observable
among rare and exclusive tulip bulbs but, with a delay, among
ordinary ones as well as the shares of the two international trading
companies. This co-movement of speculative assets cannot be
analyzed for lack of detailed data but the phenomenon is identifiable
in the other examples to be discussed. Shiller (2000) mentions an
accompanying feature of the tulip bulb speculation, which was the
increasing number of articles and commentaries in the newspapers
starting to circulate at that time in the Netherlands. The opinion
forming impact of the media cannot be considered as the direct
cause of speculation but it did contribute to attracting new investors
to the opportunity of making money on tulip bulbs. After the
zenith of speculation, February 1637, prices of both exclusive and
ordinary tulip bulbs began to fall back to 1-5% of previous values.
An adequate time series cannot be created from the available data
because we have no pricing information from the post-crash period;
nevertheless, a traditional collapse scenario is likely.
The last parameter to be examined in the definition of a bubble

is the negative macroeconomic impact. Although the above phrase
from Kindleberger mentions the outbreak of a financial crisis only
as a possibility, we believe the definition of a bubble becomes
weightless without this criterion. This is confirmed by related
literature, which touches the subject always in context of negative

effects. In the case of the tulip bulb speculation, however, this
condition is hard to verify. Some of the literature (Kindleberger
2000; Galbraith 1994) suggests that the Netherlands have been
struck by an economic crisis or at least economic decline from
the middle of the 1600s because of the tulip bulb speculation.
Although there are not much reliable contemporary data to speak
of. It is likely that the economic growth of the Netherlands indeed
slowed down resulting mainly from the depression in trading
activity. However, there are no hard facts supporting the idea
that this was the direct or indirect consequence of the tulip bulb
speculation. The previously mentioned Thirty-year War, which
ended in 1648, could play a role as well as the plague epidemic
killing about 15 percentage of the population. We cannot ignore
the fact that it was actually not a decline, only a slowdown and a
gradual falling behind England, which was to become the leader
in maritime commerce. Furthermore, Garber (1990, 2000)
emphasizes that the only primary source used in the literature is
the collection named Conversations of Gaergoedt and Waermont,
published in 1637, which was a pamphlet with a moral tendency.
In line with the intentions of the government, the writing pilloried
speculation, particularly future contracts, with an emphasis on
providing a moral to the story. The source of the previously
mentioned sailor-story is a log written seventy years after the events
whose references are of limited trustworthiness.
The most often quoted case of speculation in economic history:

the Dutch tulip bulb is put down by most economic analyses as
a craze, mad or irrational investor behavior. The above argument,
however, shows that the parallel drawn with capital market assets
is questionable to begin with, because it is impossible to specify
a fundamental value for those bulbs. The explanatory power of

the example is further weakened by the fact that no complete
data are available about the event and research was based on one-
sided, moralizing sources. Of the five Kindleberger-criteria,
negative macroeconomic impact is the most controversial to prove,
primarily for lack of information of adequate quality and quantity.
We must note that the macroeconomic consequences of non-stock-
market speculation similar to the tulip bulb one are difficult to
substantiate. Secondary trading is practically a zero-sum game
where one’s loss is another one’s gain. This basically means a
reallocation of wealth without an effect on its quantity. Still, prices
of the secondary stock market will influence primary market
parameters, namely corporate capital cost, which, in turn, might
influence economic growth in an unfavorable way. In the case of
tulip bulbs, however, the possible backlash hitting tulip growers
could not bring about serious macroeconomic consequences.
Another aspect of the problem is whether the uncertainty caused
by the tulip bulb speculation was affecting other markets, most
importantly the stock market? There is no definitive answer to
this question because of the lack of sufficient data.
A parallel between the Dutch tulip bulb case and capital market
assets is one to draw with a caveat but it is certainly useful for
illustrating the phenomenon of speculation. The overpricing or
irrational pricing of tulip bulbs cannot be proven but there are
signs of “dangerous” speculation and trading here, like the increased
risk resulting from leverage or the conceivable co-movement of
different speculative assets. We cannot analyze the impact of
newspaper articles of the time, still we can assume they had played
an indirect role in attracting masses to speculation. However,
negative macroeconomic impact of speculation, the most important
bubble parameter, cannot be verified in this case.
1.3.2 The First Stock Market Bubbles: Mississippi and South

Sea Companies
The first large-scale stock market speculations of a new era of
economic history, capitalism based on free-for-all entrepreneurship,
were happening simultaneously in France and England: those
regarding the French Mississippi Company and the English South
Sea Company. These cases, described as the first major stock market
bubbles, are characterized by the emergence of three new factors.
The first new factor was, as opposed to earlier cases, the wide
availability of issued shares. Secondly, state debt financing lay in
the background of both enterprises. Thirdly, these bubbles were
strongly connected to government activity and signs of fraud and
deception were recognizable in both of them.
Instrumental behind the phenomenon known as the Mississippi-

bubble was John Law, for whom this enterprise was the practical
implementation of one of his financial theories. Law founded Banque
General in 1716, winning the right to issue bank notes. This was
secured theoretically by the gold and silver reserves of the bank,
the volume of which was, however, much smaller than the nominal
value of bank notes issued. These notes were freely convertible to
gold and as the bank and Law enjoyed general trust, these paper
notes became widely popular. Trust was fuelled mainly by the
prospects of Compagnie d’Occident, a company law was founded in
1717. This company, known as the Mississippi Company in
economic history, was based originally on the monopoly of trading
with the French colony Louisiana and the trade in Canadian beaver.
It was renamed in 1719 to be Compagnie des Indes after the French
East-India, Singalese and China companies were merged into it.
The company thus became the sole agent of French trade outside
Europe. In the meantime, Banque General was nationalized as
Banque Royale and from then on bank note convertibility was
guaranteed by the state. This, however, did not mean increased

security for investors. Investor trust and optimism was strengthened
only by a closer tie between the bank and the Mississippi company.
The company continued to widen the scope of its activities. For
example, it has bought the exclusive rights to exploit the precious
metals of Louisiana. Later it acquired the rights to collect direct
and indirect taxes, where law saw profit potential in a more efficient
arrangement of the process. The capital necessary for growth, more
precisely the price of financial licenses, was raised from continuous
issues of increasing price. The issue price of company shares
increased tenfold from the initial 1000 livres tournois by the end of
1719. Demand for the shares was huge, fuelled by not only the
exploitation of gold reserves but also the upsurge in commercial
activities and agricultural development in Louisiana.
Money from the issue was, however, almost entirely financing a

huge state debt instead of the economic activity of the company, as
Banque Royale increased the emission of bank notes at the same
time. The foundation for Law’s financial system was the assumption
that future dividend coming from the activities of the Mississippi
Company would in the long run exceed interest payment on state
debt. This meant investment for dividend income, i.e., possessing
company shares in a long-run was “rational” – and actually, investors
interested only in dividend income were considered by Law to
rational and he was against speculative share purchases. As we have
seen in Chapter 1.1, distinguishing between shareholders on the
grounds of dividend or capital gain motivation leads to false
conclusions. In case of a public company, only a limited circle of
owners can be considered to be interested solely in dividend income,
such as the founders, strategic partners and large professional
investors. A majority of the owners can, however, be expected to
sell their shares at any time. What is more, the economic activities
ensuring profit for the company were not stable, either. Limited
information was available as to the actual economic or agricultural
potential of Louisiana and the existence of large gold reserves was
not verified.
A stock selling surge began at the end of 1719 and investor

started to demand gold for their bank notes. The regent exercising
the infant ruler’s rights conceded full power to Law over state
finances. Law was criticizing sellers as “ignorant” and he made
administrative measures, e.g., restricting the conversion of high
nominal value livres to gold, as well as other, quite unorthodox
ones. Galbraith (1994) and Strathern (2001) quotes a story where,
in order to regain investor trust, Law summoned all Parisian beggars
and marched them through the streets of Paris dressed as diggers
setting out to Louisiana. Kindleberger (2000) believes the
Mississippi Company is not a case of fraud, but is rather the result
of Law’s wrong financial theory. The question obviously lingers, to
what degree was Law aware of the workings of the company, can
we speak of deception? Although Garber (1990, 2000) and
Kindleberger (2000) identify excessive investor optimism and trust
(i.e., a natural tendency to speculate) as the main reason and they
argue Law was involved mainly in financing the company, all this
does not absolve him of the moral responsibility for deceiving
investors, which, depending on the definition, can be considered
as fraud. The financial-economic system created by Law reached
its zenith in 1720 and in May of that year the livre tournois had to
be devalued; its gold-denominated value lowered due to inadequate
cover. The share price of the company started to nosedive as a
consequence of this, by 70 percent until the end of the year. By
September 1721 the price fell to 500 livres tournois which was half
the original issue price.
The scenario for bubble formation and collapse, serving as a

framework for our analysis, was now practically completed. First, a
positive displacement took place, caused by the expectations about
Louisiana’s economic potential, the profit of international
monopolistic commercial rights and the benefits of more efficient
tax collecting in France. Secondly, new investors were drawn into
the company. The third parameter is also there: the dangers of an
“excessive degree” of speculation, not motivated by dividend income,
were signaled by leverage, which was, in this case, the financing of
state debt. Even investor deception belongs here as an attempt at
preserving trust. We can also identify a phenomenon similar to the
co-movement of financial and speculative assets seen in the Dutch
tulip bulb speculation. The value of livre tournois issued by Banque
Royale, i.e., the trust in its convertibility into gold, was closely
connected to the price evolution of Mississippi shares. Excessive
monetary expansion, the issue of bank notes without a gold backing,
caused a multiplication of the price of food, houses and real estate
in Paris (Kindleberger 2000), and this can be attributed to the
financial system created by Law, based on mutual interdependence.
The negative macroeconomic impact following the collapse of

the Mississippi Company allows, however, for different
interpretations. Kindleberger (2000) stresses direct effects, like a
general inflationary rise, an increase in state debt and a contraction
of enterprise opportunities caused by the lack of trust. Garber
(2000) believes, however, that the overall impact of the Mississippi
speculation was clearly positive as it revived the French economy
through financial innovation and the reform of state finances.
Neither of these arguments can be unequivocally verified – there
were probably positive and negative effects both in the short and
the long run. The contradiction between these points of view arises
presumably from the different weighting of speculation and the
basic characteristics of the stock market. The enterprises started by

the Mississippi Company, primarily an attempt at more efficient
tax collection, undoubtedly had a positive effect. Even “testing”
the joint-stock format could be considered as a long-term benefit.
Investing in a new enterprise based on the promise of future

success is speculation itself, which, if proving to be a failure, can
cause other enterprises to stall. Speculation may also prompt
investors draw their conclusions about the consequences of
overconcentrated, interrelated and non-transparent economic
influence (Garber 2000), which can be considered as a “proof of
sobriety”. It cannot be confirmed that French economy entered a
crisis because of the Mississippi bubble; in fact, the collapse could
just as well point out the weaknesses of a rigid economic system.
The opinion of contemporary reports (see Kindleberger 2000)is
similar to the moralizing seen during the tulip bulb speculation
and no factual conclusion can be drawn from such statements with
regard to economic or social impact.
Simultaneously with the speculation in Mississippi shares, a

speculation in the shares of the South Sea Company took place in
England, in many ways similar to the above discussed issues. The
company was founded in 1711, headed by John Blunt, primarily
with the aim of taking over state debt accumulated during the war
of the Spanish Succession. This technique was not unique in the
history of England because the Bank of England and the New East
India Company were founded in a similar form and with the same
intent in 1694 and 1698, respectively (Neal 1992). The company
was entitled to a 6 percent interest and a share issuance right in
return for the debt takeover. Its business was more transparent and
simple compared to the Mississippi Company; it was limited to
trading with the American Spanish (collectively named South Sea)
territories. In practice, this meant one commercial trip every year

as the trade monopoly was reserved for Spain.
Still, by the end of 1719 the Company managed to take over

one-fifth of the state debt amounting to 50 million pounds. At the
beginning of 1720, it has acquired the rights to manage another
31 million pounds of debt, because its 7,5 million pound bid was
better than the offer from Bank of England. The company made a
series of large volume share emissions, in the process of which share
price went up tenfold from 100 pounds by July of that year.
The popularity of the shares of the company was undiminished

in spite of the fact that it conducted no actual economic activity.
This “blind” trust was fuelled by the fact that many stakeholders
of the company had a seat in Parliament, who, with the help of
John Blunt’s informal connections and slush funds, promoted South
Sea shares while receiving share issuance rights. Such funds paid to
members of parliament and influential officials constituted the
single biggest expense of the company exceeding 2 million pounds
(Garber 1990). The success of the share issues of June and August
1720 was partly attributable to the possibility of leverage:
subscribers did not have to put up cash for the entire amount but
pay the remainder (80-90 percent) in 6-month installments.
Trade in the shares issued by the company was suspended for

the time of dividend payment due in the summer. During this
period, on August 18, 1720, Parliament ratified the Bubble Act
restricting the foundation and expansion of joint-stock companies
similar to the South Sea Company (Neal 1992). Although
apparently the Act was designed to stop the spread of enterprises
based on dubious ideas, qualified as “bubbles”, most of the literature
agree that its real objective was to ensure other enterprises, whose
spiking share price also began to attract investors, not draw
speculative capital from the South Sea Company.
Following the next capital issue, trading was resumed with the
old shares on August 31 and the price fell from 775 pounds to 290
during the next month. Investors lost faith in the South Sea
Company and an increasing number of investors liquidated their
positions, perhaps due to international events, among them being
the collapse of the Mississippi Company. By December, shares were
only worth 140 pounds. The Bubble Act, which filtered out
enterprises based on quixotic ideas (like building a perpetuum
mobile or turning mercury into a forgeable metal), also restricted
the foundation and functioning of other companies created in a
joint-stock format up until 1824. This can probably be considered
as the long-term, harmful economic consequence of the case.
The two previously discussed examples presented basically all

parameters of a stock market bubble, but the weakness of the
Kindleberger-framework, the criterion regarding speculation, was
again highlighted. We can calculate that the highest market value
for the South Sea Company was 164 million pounds. If we deduct
interest payment in return for state debt takeover and cash claims
from postponed subscription payment, we are left with
approximately 60 million pounds, the cover for which would have
been the profit on future business (Garber 1990, 2000). As the
South Sea Company did not conduct actual economic activity, we
can speak of pure speculation for trading gains. However, when a
company conducts actual business, like the Mississippi Company,
such buyers cannot be distinguished from dividend-oriented
investors. In such cases a price decline is not the result of
overvaluation, rather a change in investor trust, and due to a natural
feature of secondary markets, restricted liquidity, a collapse is
inevitable.
Regulation similar to the Bubble Act of 1720, trying to weed

out “unstable” companies, is practically impossible to draw up
because “there is no basis for distinguishing bubbles from other

types of financial assets” (Gilles and LeRoy 1992:74).
Hereinafter we are not going to focus on speculation per se, rather

on parameters which may not causally imply bubble formation or
bursting but which may signal a change in the confidence, share-
buying attitude and capacities of market players. These are: leverage,
co-movement of asset prices, state intervention and managerial
action aimed at investor deception.
1.3.3 The Crashes of 1929 and 1987

A recurring phenomenon of stock markets is when investor attention
turns to specific companies or industries, widespread speculation
ensues, and the process runs full circle with a significant price
decline. Of such declines, the US stock market crashes of 1929
and 1987, and the so-called Dotcom speculation connected to the
latest technological revolution are the examples remembered most
vividly in economic circles. The following two chapters will present
the characteristics of these cases, and the contradictions of their
economic evaluation.
The US economy experienced technological and structural changes

in the middle of the 1920s. American companies started to adopt
mass production, manufacturing a broad scale of products with
increasing efficiency. Consumer habits were changing, too. More
and more inventions took their place domestically and electricity
started its conquest of the US cities and homes (Shiller 2000). A
wave of corporate capital issue began, and share prices in the most
influential US capital market, Wall Street, started to surge. Already
palpable investor optimism found new momentum after the Florida
real estate boom and the subsequent huge price decline, which
sent a new army of speculators to stock markets. New financial
enterprises emerged: firms with professional investment activity

also involved in capital market consultancy and share trading; and
companies bought shares and issued shares to refinance
(Kindleberger 2000). The first were the predecessors of today’s
investment bank, the second were the forerunners to modern closed-
end funds. Popular types of financial transactions included
investment formats where private individuals borrowed against
shares bought from the loan. In some cases, a 10 percent margin
was enough for a stock purchase with the rest lent by a bank. The
sale constraint caused by leverage was therefore present in the
decision-making of market players, leading to a shortened
investment period, increasing the “danger” of a potential large-
scale price fall.
Share prices soared almost universally throughout 1928. In

March 1929 the Federal Reserve Board hiked rates but this caused
only a minor slowdown in an upward trend. Still, it became
increasingly clear that the government actively attempts to put a
brake on share prices. Friedman and Schwartz (1963) identifies
this mistake of monetary policy as the cause of the eventual collapse.
Galbraith (1994), however, compares the growing stock market
speculation to mass delusion, coming to an end in the autumn of
1929. There was an average price decline of 12.34 percent on
October 28, followed by another 10.16 percent the next day and
9.92 percent on November 6 (Schwert 1992. p.577). The crash
ending stock market speculation for years to come was underway.
The US stock market speculation of the 1920s displays several

characteristics of a bubble. Liquidity expansion resulting from
leverage was not necessarily a driving force behind the boom because
only 5 percent of all public shares were financed from loans
(Cecchetti 1992). Broker credits required a 200 percent cover
(Cecchetti 1992), although sale constraint was often an additional

factor in a price decline due to smaller margins. Leverage produced
by broker credits added to the number of future sale offers but this
involved only a fraction of the entire stock volume suggesting it
was not the most important factor behind the collapse. In the case
of banks offering financial services, the moral hazard presented by
the action of investment agents could contribute to the crash.
Galbraith (1994) brings up the examples of United Founders

Corporation and Goldman Sachs Trading Corporation, where a
shadow of suspicion might arise about inadequately informing
investors of transactions and the effect and risks of leverage. The
share price of these funds was 30 percent higher than the value of
financial investments providing the bulk of assets. Still, this alone
does not prove that public companies in general were overvalued.
White (1990) analyzed corporate quarterly dividend data and could
find no proof of “artificial” share price influencing on the part of
managers by pumping up dividends. Signs of large-scale speculation
do pop up but this does not necessarily imply unrealistic prices.
The output and profitability of companies did not decline in this
period (Friedman and Schwartz 1963) though the growth rate of
dividend payment was lagging behind the rate of share price rise.
This, however, was more a result of managerial pessimism than
deteriorating economic outlook (White 1990). Technological and
structural changes influenced investor expectations in a positive
way. The crash was primarily not a consequence of deteriorating
fundamentals. The 30 percent devaluation resulting from the share
selling surge can even be deemed a negative bubble (Shiller 2000).
According to Cechetti (1992) and Friedman and Schwartz (1963)

price decline was due to government intervention, the shortcomings
of monetary policy. The Federal Reserve Board increased interest rates
three times during 1928, with brokerage credits increasing

concurrently. Risk premium attached to brokerage credits was
fluctuating but here, too, an upward trend is recognizable implying
that credit expansion was not a significant driving force behind capital
market rise (White 1990). Whether intervention from the Federal
Reserve Board was necessary continues to be a subject of debate. It is
a fact that not only monetary authorities but also the government
was committed to act against stock market phenomena seen as
negative. Cechetti (1992) quotes the following note from the memoirs
of Herbert Hoover, was elected as the US president during the month
of March of that year: “To create a spirit of caution in the public, I
sent individually for editors and publishers of major newspapers and
magazines and requested them systematically to warn the country
about the unduly high price of stocks.” Besides direct influence,
therefore, the government tried indirect methods to cool stock market
speculation. Of course, the media published not only a negative
outlook but also articles promoting the stock market, starting columns
on the subject, quoting experts forecasting a lasting boom. Galbraith
(1994) and Sornette (2003) quote the case of Irving Fisher, a respected
Yale professor, who said 14 days before the crash: “In a few months,
I expect to see the stock market much higher than today”. Analyzing
news from October 1929, Shiller (2000) could not find an explanation
to what the direct cause of the collapse might have been.
Regardless of the cause, the US stock market indices kept on

falling until 1932, losing, on average, 70 percent of their value.
The 1929 crash ushered in the Great Depression hitting not only
the US but world economy in general. The US production was down
50 percent compared to pre-collapse levels. The stock market crash
undoubtedly had negative macroeconomic consequences, mainly
in the form of increasing income insecurity. Cechetti (1992) states
the demand for non-perishable consumer goods was dramatically
down as a result of the stock market crash and consumer

indebtedness dating back to before the collapse. The latter was
directly affected by the crash as the Federal Reserve Board did not
give up its strict monetary policy curbing demand in the long run.
Another result of the crash was a transformation of financial
regulations. The Glass-Steagall Act ratified in 1932 separated classic
banking activities from investment banking and the Security
Exchange Commission was formed.
While many see the 1929 crash as the direct result of doggedly
restrictive monetary decisions on part of the Federal Reserve Board,
58 years later another famous collapse, the 1987 crash, was traced
back to altogether different factors. The economic background
behind the boom of the 1980s was Ronald Reagan’s tax-cutting
economic policy which induced growing consumption and
increasing corporate output (Soros 1994). The engine of economic
growth was the government spending and foreign capital flocking
to the US. The corporate sector was flourishing: a merger wave
began, encouraged by favorable regulatory conditions. Junk bonds,
the name referring to an especially high risk attached, became
widespread as a capital market financing method, mainly for
leveraged buy-outs. This was helped by the possibility of tax relief
for corporate bonds, making buy-outs easier and the number of
potential targets higher (Schwert 1992). Stock market prices rose
sharply beginning in the middle of the decade, in line with corporate
takeovers. Trading volume went up as well, while real estate prices
also started to climb. Although Galbraith (1994) in a 1987 article
thought an inevitable collapse was predictable, there is no clear,
widely accepted explanation for the direct cause of the crash.
On “Black Monday”, October 19, 1987, the Dow Jones

Industrial Average (DJIA) fell by 23.1 percent while the Standard
& Poor’s Composite Index (S & P 500) was down 20.4 percent.
A survey of both institutional and private investors prepared by

Shiller (2000) stressed “market overvaluation”. Average price per
dividend (P/D) and average price per earnings (P/E) reached earlier
historical highs recorded in the 1920s, 1930s and 1960s (Schwert
1992). The 1960s boom was, however, not followed by such a
drastic price decline, implying other reasons behind this sudden
fall (Sornette 2003). So-called programmed trading may have
played a role in the large number of sales and record-high trading
volume. This type of trading, widely applied ever since, means
bids and complex positions are fed into computers and completed
automatically in line with conditions set in advance. This method
for investors to avoid further losses was rated important in the
Shiller-survey. An example for complex positions is portfolio
insurance, whereby securities were hedged by short put options or
short future sales. So-called stop loss commissions were also
spreading which resulted in an automatic sale order when the price
of a given share dropped to a pre-specified level. On the day of the
collapse, the trading system could not manage large volume sale
orders in the prompt market, while futures liquidation happened
faster creating losses in the process for portfolio insurance through
basis risk.
Two pieces of economic news merit a mention with regards to

the crash. Articles appeared in the preceding days about the US
budget and trade deficits and their negative economic implications,
and these were emphasized by institutional and private investors
in the Shiller-survey. Trade deficit hit a new low in October 1987
not seen in a decade, which prompted speculation against the dollar
(Soros 1994). Schwert (1992) finds a weakness in this argument:
in coincidence with the decline of DJIA and S & P 500, almost
every significant stock index started to fall. The fundamental reason
being an “excessive” capital inflow bearing a negative impact on
the US economy does not stand, because this may even be favorable
for other countries. An important piece of news could have been
the admonition by Robert Prechter, an investor known as a “market
guru”, who prompted others to sell. The New York Times published
a price diagram comparing the current boom to that of 1929, raising
the specter of a similar collapse, which indeed happened just a
couple of hours after the paper went into circulation. The literature
is unanimous in concluding that this crash had no long-term
negative macroeconomic impact whatsoever and it was not followed
by a general liquidity crisis. The Federal Reserve Board relaxed its
previous monetary policy, not committing the mistake of 1929.
Still, both the US and international stock markets were characterized
by uncertainty, and changes in economic policy were inevitable.
A commission was formed, headed by Nicholas F Brady, to

identify the reasons behind the stock market crash. The report
prepared by the commission listed all the above-mentioned
technical reasons, highlighting portfolio insurance and programmed
trading, but it failed to mention other factors affecting stock market
speculation. As a result of the work of the Brady-commission,
regulatory changes took effect, the most important of which was
the possibility of trade suspension that automatically applied in
the case of extreme price fluctuations. The practical benefit of such
trading-related regulation is debatable: to throw “sand” in the
wheels of trading to slow down investor reaction is not supportable
from the side of theory. Closing a stock market or suspending papers
does not influence the degree of fluctuation, and bubbles might
form even given high transaction costs, like in the case of the real
estate market (Shiller 2000).
Elements of the Kindleberger-definition of bubbles can be

identified in the stock market boom of the 1980s but examining
additional features gives a more precise picture of the circumstances

of speculation. Besides leverage, discussed as the most important
cause by the Brady-report, we cannot lose sight of the impact of
governmental economic policy. In contrast to the active fiscal policy
of the Reagan era strongly stimulating both supply and demand,
as part of its monetary policy, the Federal Reserve Board began to
pursue a stricter interest policy from the summer of 1987. This
Janus-faced economic policy had, therefore, both positive and
negative impact on the stock market, and news on the budget and
trade deficit strongly influenced market prices as well as the high
interest rate strengthening the dollar. These could play an important
role in the crash by orienting and forming investor reaction as
external factors. Another characteristic was the high moral hazard
carried by the actions of managers in charge of corporate buy-outs.
In some cases one could speak of fraud and inadequate disclosure
(i.e., investor deception), which were verified afterwards (Galbraith
1994; Kostolany 1991).
The stock market crash of 1987 is interpreted as the end of a

positive bubble, where the prices of previously overvalued shares
reverted to their economically correct level, or as the beginning of
a negative bubble, where uncertainty that resulted from panic caused
shares to trade temporarily below their realistic value. Although
these two definitions allow for different explanations, we believe
neither can be fully confirmed. As soon as we qualified the term
“bubble”, further interpretation problems would arise. In our
opinion, therefore, the understanding of the bubble phenomenon
is aided by the identification of the common roots of the
above-discussed stock market collapses.
Putting aside terminological issues, we can be sure of one thing:

investor opinions took a huge turn in the course of a couple of
days. The above-mentioned indications might provide the reason

for a sudden change in speculator attitudes but we can confirm
this only in hindsight, i.e., when the market has already collapsed.
1.3.4 Internet Bubble

The last great stock market boom in economic history, lasting from
the middle of the 1990s until the millennium, was hallmarked by
the rise of the Internet. This “Internet bubble” burst in 2000,
bringing about the crisis of the infotech-sector. Yet again,
emphasizing the irrational overvaluation of share offers an inadequate
explanation being too vague and hard to support by facts. The
large scale rise and sudden fall of information technology,
communications and Internet shares can be tied to several different
factors. In the following paragraphs, this bubble phenomenon will
be analyzed focusing on Internet-related IT-firms.
Having overcome the recession of the early 1990s, the US

economy began to grow. General optimism was helped by political
events, too: the end of the Soviet Union, the conclusion of the Gulf
crisis, and a new era of European unity. Due to the moderate rate
of inflation, investor confidence began to climb back with regards
to shares following the unpleasant memories of the junk-bond
fuelled wave of corporate buy-outs and the Mexican credit crisis of
1995. In addition to this, from the middle of the decade the so-
called baby boom started to make its impact. This term refers to
the postwar generation causing a demographic boom. Members of
this generation entered their 40s and 50s by the end of the 1990s
and their saving and spending habits became decisive factors both
in stock markets and the markets of products and services. Following
the low of 1794.6 points after the 1987 crash, in eight years the
DJIA more than doubled to 3839 points (3 January, 1995),
similarly to NASDAQ (National Association of Securities Dealers

Automated Quotation) which comprised small-cap shares as well,
going up to 744 points.
A drastic change was under way from the middle of the 1990s in
the corporate, governmental and domestic use of information
technology. The market for personal computers experienced dynamic
growth; new companies appeared at the side of long-established IBM,
becoming serious contenders in many market segments. A milestone
in software development was marked by the launch of the Windows
95 operating system, and later the Internet Explorer, by Microsoft.
Corporate application fields for computers also expanded to include
internal networks, databases and integrated planning and controlling
systems. Companies went head over heels to improve their IT-systems
creating strong demand for both hardware and software. This
prompted intensive development in such products and by the
millennium, the performance of microprocessors, computers and
peripheries rose by several orders of magnitude as their price fell to
the fraction of previous levels. The other factors influencing the market
was the Internet. Besides earlier popular tools (gopher, ftp, e-mail),
the emergence of the World Wide Web (WWW), a convenient,
graphics-heavy hyperlinked pages brought along qualitative changes.
The Internet provided a new channel for companies to their consumers
and clients. Full return on related spending was, however, only to be
expected in the long run, with initial investment eating up huge
amounts of money.
The IT-revolution made its symbolic debut in the stock market

in August 1995, when Netscape, the developer of the then market
leader browser, was listed. Netscape-shares rocketed from the issue
price of USD-7 to 36 on the first day of trading. Later it came back
to lower levels but by the end of the year it was traded at almost
100 dollars. 1996 saw a large wave of capital issues, mainly related
to the creation of so-called dotcom-companies with an Internet-
oriented business model. During initial public offerings, a wide
scale of investors had, for the first time, the opportunity to buy a
stake in such companies, providing fuel for later trading. In these
days it was not uncommon to see share prices multiply on their
first day on the floor. In the US capital markets, there were 40
huge IPOs (see details at http://www.ipoinfo.com).
Already established IT and communications companies also

began to capitalize on the desire of investors to buy into the new
industry. These companies used new capital issues to finance the
acquisition of smaller dotcom-firms. Another doubling in the value
of stock indices took only two years from 1995. After the Russian
and Far East crises at the beginning of 1999, the growth rate of
indices changed: DJIA rose by another 30 percent in subsequent
years, while NASDAQ almost quadrupled during the same period.
DJIA reached its maximum on January 14, 2000 at 11.723, and
NASDAQ peaked out at 5048.62 on March 10, 2000.
Figure 1.3 shows that these stock indices returned to their 1998
levels by the autumn of 2001, which was still 250 percent higher
than in 1990. The rise and fall of dotcom-shares was more drastic.
The DOT index, calculated from November 13, 1998, representing
the share price evolvement of Internet-related companies, surged
from the initial 254 points to 1333 on March 9, 2000, which is a
more than five-fold increase. By September, 2001, it was a bit more
than one-tenth its peak value, a mere 142 points.
The elements of Kindleberger’s bubble scenario are identifiable

here, too. The economic push was provided by the increasing role
of IT both at the corporate and domestic level as well as by the
Figure 1.3: DJIA, NASDAQ and DOT between 1995 and 2001
2400%
2200%
2000% DJIA
DJIA
1800%
1600% NASDAQ
NA SDA Q
1400% DOT
DOT
1200%
1000%
800%
600%
400%
200%
0%
Jan-95
Jan-96
Jan-97
Jan-98
Jan-99
Jan-01
Jul-95
Jul-96
Jul-97
Jul-98
Jul-99
Jul-00
Jul-01
Note: For DJIA and NASDAQ 01.02.90 = 100% and for DOT 11.13.98 = 100%
Source: own calculations based on data from finance.yahoo.com
emergence of the Internet as a new communication channel.

Countries utilizing the benefits provided by these opportunities
experienced a huge rise in efficiency. The drawback of all this was
the vast capital expenses incurred during the replacement of
corporate PC-arsenals, the implementation of the hardware and
software infrastructure for online sales and purchases and the
acquisition of “one-idea” dotcom firms. A new industry of creating,
managing and transferring information was the key to efficiency
growth, widely referred to as information economy, e-economy or
new economy. Although these terms do not exactly cover the same
phenomenon, and it is still a matter of debate whether we can
speak of radical changes in economic processes or even in the science
of economics, the positive impact on economy is indisputable.
Bõgel’s (2002) analysis supports the view that the US economic
boom and the improvement of its economic indicators (productivity,
inflation, unemployment rate) were clearly linked to the emergence
of new technologies. New players were coming from the ranks of

the baby boom generation, who increased the proportion of stock
investment within their portfolios. This is illustrated by the fact
that while in 1984 only 20 percent of the US households possessed
shares, in 2000 that number was 27.1 percent. The proportion of
shares in the net wealth of households went from 6.8 percent up to
15.6 percent due to rise in volume and prices between 1984 and
2000, with its maximum reached in 1998 at 18.8 percent
(see Figure 1.4).
Figure 1.4: Changes in Share Possession of US Households

30.0% ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
25.0% ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
20.0% ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
15.0% ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
10.0% ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
5.0% ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
0.0%
1984 1988 1991 1993 1998 2000
Note: Lighter column = proportion of shareholder households.
Darker column = proportion of shares within net household assets.
Souce: US Department of Commerce (www.census.org).
The NASDAQ, which served as a low transaction cost automated

secondary market for the US mid-cap shares, played an important
role in facilitating the appearance of companies with large growth
potential founded in the 1990s. It also provided cheap and easy
access to shares for a large number of investors. Thus, the Internet
was not only a germinating force for speculation but also a technical
stimulant of stock trading through lowering transaction costs. This
also led to the transformation of services offered by brokerage firms
and investment banks, because the consultancy role of such

companies dwindled due to a warehouse of free investment
information and analysis available on the Internet, but at the same
time they advanced their online trading facilities (Barber and Odean
2001; Komáromi 2002b).
When a new technology emerges, it is always difficult to predict

the future cash flow of a company, the risks of innovation, and the
expected yield derived from this, in other words, the fundamental
value of a share. An example of valuation uncertainty is the famous
statement made by Alan Greenspan, the head of the Federal Reserve
Board, in a conference of the American Enterprise Institute for
Public Policy on December 5, 1996, in which he used the term
“irrational exuberance” to describe stock market events. At the time
DJIA was at 55 percent of its future maximum with 6437 points,
NASDAQ was at around 25 percent of its subsequent peak with
1300 points. Analyses in The Economist began to refer to stock
market bubbles a couple of months before the historical maximum
of these indices was reached (The Economist 1999a-b) comparing
the boom of technology shares to famous earlier speculations,
showing diagrams of price rise and fall. The upward trend of DJIA
and NASDAQ nevertheless continued unimpeded; in March 2000,
the indices were still going up.
A mixed picture is painted about what corporate or

macroeconomic news formed investor reactions at the time. On
March 6, 2000, at the Boston College conference on “New
Economy”, Greenspan addressed issues like structural discrepancies
in macro-level supply and demand, an overstretched labor market
(unemployment hit a 30-year low), inflationary threat and
imbalances in the economy. The DJIA, and a day later NASDAQ
and DOT reacted to the news with a decline, but, like many times
before, this decline was temporary. Favorable US productivity data

were published on March 7, causing the latter two indices to rise
to historical highs. DJIA, the representative of the shares of
traditional companies (or to use an already widespread term, the
“old economy”), was held back by the profit made by the
management of Procter and Gamble. Two days later the business
media reported temporary profit-taking and a potential trend
change after the decline of share prices in the Japanese technological
sector. The majority of international stock market analysts forecast
further price rise and further divergence between DJIA and
NASDAQ. Nikkei 225, the Japanese stock index set a 40-month
record and the Beige Book reported robust growth in the US
economy. DJIA and NASDAQ were a couple of percentage points
down with high volatility.
The antitrust case against Microsoft, potentially unfavorable to

the company, was followed closely in the IT-communications sector.
This may have changed a role in an apparent change in sentiment,
with investors starting to favor old economy shares once again. The
media kept uncovering cases of “creative accounting”, irregularities
in running the books of companies, prompting inquiries. While
NASDAQ lost 10 percent of its value in the course of a couple of
days, DJIA kept rising to new highs. The Federal Reserve Board
hiked the base rate by 25 percentage points on March 21, to reach
6 percent. Usual quarter-end portfolio reshufflement resulted in
further price decline. At the beginning of April, IT-communications
sector shares suddenly began to fall both in the US and Europe.
NASDAQ dropped 7.6 percent on April 3, losing another 5 percent
the next day and falling yet another 10 percent on April 16 to hit
3677 points. The rest of the year saw strong volatility, excessive
daily rises and declines in the value of NASDAQ. The trend was
clearly downward now and by September 10, 2001, the index
reached 1,695 points, 27 percent of its historic high. DOT was

down at 142 points, 11 percent of its peak value.
Important reasons behind the rise and fall of technology shares

included not only capital market instability but also institutional
factors, market deficiencies and the influence of government policy,
forcing IT-communication companies into a monopolistic
competition, and thus, drastically squeezing profits, as highlighted
by Major’s (2003) study. The study brings up examples of
overpriced government licences in the telecommunications sector,
the purchase of which stretched the finances of companies. With
the whole sector slowing down, these businesses inevitably went
into the red. Analyzing the IPOs of dotcom companies, Bogan
(2003) concluded that high initial quotation was largely due to
the specificities of ownership structure. The appearance of
institutional investors pushed priced up but this cannot be
considered as irrational pricing.
Investor behavior in the secondary market was often characterized

as irrational. However, this cannot be clearly supported, either.
The average P/E of S & P500 shares, calculated by Shiller (2000),
was 44.3, 30-40 percent over previous highs recorded in 1929 and
1966. Current P/E is in a weak negative correlation to later yields,
which means a high P/E forecasts a low or negative average yield in
the stock market. To support the argument for overpricing, we could
mention that at the millennium average, dividend level was very
low in light of historical data. Can we reasonably state that investing
in an average portfolio in 2000 was an irrational act? If we consider
the correlation between P/E and future yield to be statistically
significant, we are faced with the methodological problem of what
resemblance the past bears to the future. Negative correlation is
valid only in case of unaltered investor behavior, e.g., a constant
stock proportion within financial assets kept. This indicator was,

however, significantly changing, as shown in Figure 1.4.
Throughout the 1990s, investors were putting an increasing share
of their financial assets in stocks, with the enormous savings of the
baby boom generation further raising demand for stocks.
A 40+, value of P/E might therefore seem high in historical

perspective but in modified investment habits, preference for riskier
securities and a stepped-up demand could well explain this level of
the indicator. The usefulness of P/E is limited by the fact that
profits were zero or negative, especially so in the case of dotcom
firms, which strongly dragged average profit per share down. This
amplified the natural flaw of averaging, concealing the distribution
of individual values. P/E has no limiting value. We can find rules of
thumb for expected value depending on the type of share, but this
benchmark is impossible to specify in the case of a new, qualitatively
different, non-comparable industry. Calculating with 1998 and
1999 results, dotcom shares were characterized by P/E values
between 50 and 200. Shiller’s (2000) P/E analysis does not infer
irrational behavior on part of stock market investors of the
millennium. A high P/E might simply indicate that long-term
buyers become more scarce, with a dominance of short-term
investors. The same can be deduced from low dividend levels as it
fails to be a magnet for “lifelong” investors.
Apart from P/E, nothing really indicates an overvaluation of

NASDAQ or DOT shares supporting the claim of irrationality.
Investing in companies with zero or negative results does not mean
an irrational behavior. An investor is not interested in current
corporate profits but expecting later profits to be high. The buyer
therefore “makes a bet” on the profitability of the industry and the
given company within that. Comparing share prices for the new
and old economy do not indicate an overvaluation of the former.
For example, the market capitalization of Cisco was seven times

higher than that of Ford in 2000, but then again, those buying
Microsoft shares in 1986, the year the company was listed, saw
their investment multiplying 300 times by 2003, implying they
were successfully betting on the future of the company. Those who
bought Cisco shares near its maximum cannot be deemed irrational
investors as it is a question of years to come whether their optimistic
expectations were justified by the success of the firm. Cisco’s share
price being a fraction of its historical high means only “unfortunate
timing” on part of some buyers.
The price rise of a company stock, fuelled by investor demand,

is in many ways resembling the previously discussed Ponzi-scheme
but it can only be considered irrational if the expected dividend or
trading profit is unachievable. This is very rarely the case with stock
investments. Dividend payment is not a “must” for companies, in
contrast to interest payment on loans or bonds. In the case of most
dotcom firms, not paying dividends for many years, owners were
well aware that initial corporate investments will turn into profits
for the company and dividend for shareholders only later. The
success of investments made by the company, and thus their impact
on the fundamental value of the share (and the company), will
only be evident in hindsight. So looking at stock investments from
this angle, most of them cannot be considered irrational.
There is only one case when share price may be clearly overvalued,
and this is the case of investor deception on part of managers fully
aware of the actual situation of the company. The current situation
and future potential of a firm is best judged by the managers who
possess most of the relevant information. The relation of owners and
managers can be described with the agent-principal problem which
occurs because of different information and interests. In line with
the management principle emphasized throughout the 1990s,

maximizing shareholder value, corporate management was receiving
a large proportion of its compensation in the form of stock options,
making their income strongly dependant on share price performance.
Apparently, this solved the conflict of interest between the two groups,
both being interested in a continuous price rise. However, the agent-
principal problem did not go away, because managers were thus
prompted to manipulate corporate data if they intended to exercise
their options. This could be done by short-term methods, which
were coinciding with the interests of shareholders, but informational
asymmetry remained as owners did not know how long this level of
profitability can be maintained. Cases of “creative accounting” kept
surfacing, when managers massaged balance sheets and income
statements deceiving investors. In the owner-manager relation
characterized by asymmetrical information, dividend plays the role
of signal. Dotcom firms, however, did not pay dividend in order to
exploit growth opportunities; so remote was even the theoretical
possibility of dividend payment signaling financial stress and corporate
difficulties. The spread of stock options therefore did not solve the
problem. When, throughout the year 2000, suspicion of “creative
accounting” was raised in the case of an increasing number of
companies. This could also affect, through distrust, the shortening
of the investment period. Interestingly, most scandals involved
companies of the old economy. The explanation for this might have
been that managers of such companies were hard-pressed to keep
track of their investors who were flocking to the stocks of the new
economy urged by the rise of dotcom shares, which they did by
pumping up profit figures. Nevertheless, scandal after scandal did
contribute to a shrinking trust in shares of either old or new economy.
From the macroeconomic point of view, the stock market boom

of the 1990s was unfolding in the presence of favorable developments.
The US GDP in the second half of the decade was growing at a rate
on average 1.3 percentage point higher than before, while average
inflation was one percentage point lower than in the preceding four
years (see Bõgel 2002). Stock market price rise was attracting new
buyers to secondary markets but demand for new economy shares
was high in the primary markets, too. Internet-related shares promised
high profits for investors, significantly influencing capital allocation
in the economy. In other words, the fashionable technological sector
sucked capital from other industries. Firms receiving “risk” capital
made huge IT-investments increasing efficiency and contributing in
an important way to the US economic growth. The dominance of a
sector may be harmful, driving investors out of other industries, if
this is coupled by an interest rate hike i.e., a direct increase in corporate
capital cost. Greenspan, as the head of the US monetary policy, was
closely followed by the public ever since his previously mentioned
“irrational exuberance” speech of 1996. His public speeches and
commission appearances as well as the decisions and explanations of
the Federal Reserve Board became the most important pieces of the
US economic news (Shiller 2000). Fund interest rate was lowered
from 5.5 to 4.75 percent in 1998, then in June, August, November,
1999 and February, March, 2000, it was raised by 25 basis point
each, to reach 6 percent. Consumer inflation rate did not change
much in the meantime, and although rate hikes were routinely
accompanied by emphasizing inflationary pressures, the Federal
Reserve Board was ultimately pointing out the overvaluation of stock
markets. As we could see in the above news analysis, share prices fell
significantly in April, 2000, but another rate hike was carried out on
May 16, now by 50 basis points, increasing base rate to 6.5 percent.
Alan Greenspan was “successful” in the sense that DOT and NASDAQ
shares plunged. It is still a matter of debate whether share of some
particular sectors were falling to their “realistic” level, and whether
the burst of the bubble was indeed inevitable.
The US economic growth was 3.1, 2, 5.2 and 7.1 percent in the
quarters of 1999; 2.6, 4.9, 0.6 and 1.1 percent in 2000; however,
the first three quarters of 2001 recorded a negative GDP change
(–0.6%, –1.6%, –0.3%), which, by definition, signaled a recession.
The same was implied by the decline in real asset investments starting
from the fourth quarter of 2000, lasting two years. O’Quinn’s (2003)
analysis prepared for the US Congress Joint Economic Committee
stresses that the 2001 recession, in contrast to post-war crises
stemming from declining consumption, was the result of earlier
mal- and over-investments, for which he offered examples from
telecommunications and media sector companies. Overinvestment
was also found in traditional companies investing in e-business or
e-commerce; software and hardware developments proved precipitate.
One underlying reason was strong competition and the pressure to
outbid rivals. With Internet-firms, it was either a case of failed ideas
or lagging efficiency in production, commerce and application.
Although the rate hikes of the Federal Reserve Board were made on
the premise that overpriced shares caused excessive consumption
creating an inflationary threat, the real long-term danger was
presented by an intergeneration income reallocation and an
inappropriate capital allocation threatening economic growth (Baker
2000). When a new industry becomes dominant, as with the
technological sector at the turn of the millennium, coupled with a
universal stock market boom, conceivably positive effects will
outweigh negative ones in the long run. Dollar-billions flowing
unchecked into dotcom firms went out the window because
innovation financed by them proved to be heading nowhere. A
re-examination a couple of years after the collapse of the Internet
bubble suggests, however, that the experience of unsuccessful
innovations and investments actually did help subsequent advances
in productivity and economic growth (Bishop 2003).
The US monetary policy considered the stock market boom of

the 1990s to be a bubble, trying to prevent or at least mitigate
some of the objectively negative macroeconomic impacts. This does
not necessarily mean, however, that dotcom investments were
irrational decisions. True, investors did not interpret the message
of continuous rate hikes correctly but it was impossible to precisely
define the interest rate level which would trigger the crash. As in
the previous examples, it is also apparent in the case of the Internet
bubble that, sooner or later, the opinion of those in charge of
economic policy might prove crucial. If a capital market boom is
classified as a “bubble” by the decision-makers of economic policy,
a collapse is only a matter of time.
1.3.5 Conclusions
The conclusion made by literary economics about stock market
bubbles is that their most important feature is the crash. Economists
would hardly speak of a bubble without a collapse of prices. When
financial asset prices sink in a gradual, continuous manner, it is
presumably not a cause but a symptom of an economic recession
or decline. Besides the bubbles presented above, Kindleberger
(2000) discusses another 28 financial crises up to and including
the 1998 Russian case, which were accompanied by stock market
speculation during the last two-and-a-half centuries. Other works
on financial history (e.g., Galbraith 1994) also show that the term
“bubble” is only attached to events with negative macroeconomic
consequences or a crisis. There is no causal relationship, however,
between a crisis and the formation and collapse of a bubble. The
latter may follow or accompany an economic crisis and in some
cases a stock market crash can indisputably deepen a crisis. The
identification of a negative impact is indispensable in order to lend
weight to a bubble. A widespread speculation is basically a way of
wealth reallocation between two investing parties (winners and
losers) but the conservation of a bubble is in the interest of

intermediaries (brokerage firms and investment banks), and we may
even have to realize that a fourth party, the government, is also a
beneficiary of a stock market boom. Ways in which this presents
itself include a direct enhancement of economic performance by
the investments made or the elevation of potential output level by
innovation. In the end, all this may convert into votes in elections.
The “moment of truth” following a crash can be a stimulant for
beneficial structural changes in the economy. The most important
negative economic impact is not direct: it is the deterioration of
investor confidence. After a stock market collapse investors lower
the proportion of stock in their portfolios causing an increase in
the cost of capital for companies.
Traditional analyses, however, often fail to consider that shares

of a potentially successful new industry are bought at the beginning
of a bubble exactly because this investor confidence is “unduly”
high. Even if the benefits of stock market allocation are not clearly
supportable in comparison to institutional (banking) allocation
(Bácskai 2003), the general trend of securitization does imply that
spreading risks among a large number of investors is one such
benefit. Another advantageous feature of stock market, is the fact
that capital owners renounce their capital “for good” resulting in
less restriction and more commitment in the case of bank lending.
For innovative enterprises, or where a new industry is emerging
with the need for risk capital, a stock market does a better job at
financing because the myth of a “new economy” or investors
influenced by a “herding effect” will facilitate the capital acquisition
of trend-setting. As a negative consequence of this, mal- and over-
investment might drag down profitability in the middle run, and
run up the cost of capital when disappointed investors turn their
back on their previous darling of an industry. The economy of a
nation whose capital market is made more unstable by this may,
however, enjoy faster long-term growth because, following a period

of possible slowdown, innovative companies will make use of the
previously accumulated experience. Szalavetz (2004) believes it is
primarily not the US economy that will benefit from the Internet
bubble this way. In our view, however, there is nothing to indicate
that, recovering from capital market depression, the US economy
provides less opportunity for further innovation and IT-
communication investment than other developed or developing
countries. The commonplace statement “no pain no gain” (like in
The Economist 2003. p.74) proved to be right in several historic
cases discussed above.
Literary economics does not offer an answer to what the necessary

and sufficient conditions of a stock market bubble are. Kindleberger’s
definition aided us in assembling the typical traits of a stock market
bubble but the criterion of “excessive speculation” is inadequate for
many reasons. The analyses of stock market examples made clear
that investors, analysts and decision-makers of the economic policy
will, from time to time, consider some stock market phenomena as
“excessive speculation” and qualify them as “bubbles”. In their view,
bubbles do exist, and their behavior and reactions can even cause
the burst of such bubbles. Lacking a practical definition we cannot
provide an acceptable answer to whether, from the scientific point
of view, stock market bubbles indeed exist and whether we can
distinguish these from other stock market phenomena.
Building upon the conclusions of different economic approaches,

in the following chapter we will attempt to create an analytical
framework for the phenomena characterizing stock market
speculation, according to historical experience and empirical
research. These phenomena are indicators of an increasing
probability that prices are peaking and the main feature of a bubble,
a stock market crash, will ensue.
Anatomy of Stock Market Bubbles 69
Anatomy of Stock Market Bubbles
“We should answer for a single question: are there more idiots
than stocks, or more stocks than idiots?”
André Kostolany (1991. p.34).
) s we saw in the previous chapter, there is no entirely valid and

practically useful definition either for bubbles or asset price
bubbles. Hereinafter, we will narrow the issue down to examining
whether a change in investor behavior, sufficient for causing a stock
market crash, is recognizable.
A strict definition for a stock market bubble will not be provided

by us, either, but, with the help of factors increasing the likelihood
of a crash, we aim to give an indirect answer to when we can qualify
a stock market phenomenon as a bubble. Then we will list the
stock market booms and crashes of economic history using the
typical traits.
2.1 Investment Decisions and Noise Trading

Investor behavior can indirectly be analyzed with the help of price
fluctuations. We set out from the assumption that investors are
making their decisions not randomly but in a conscious manner,

based on some information. By information we mean any news
related to the particular company or country, result of an analysis,
or any reason prompting an investor to modify the composition of
his portfolio.
Price may change if information is symmetrical, i.e., available

for every investor and interpreted in the same way. For a change in
price it is enough, however, if only a portion of the investors receive
the news, or if they interpret its impact on share price differently.
In such cases, price changes due to asymmetrical or different
information can be divided into three types: One is information of
general importance (of interest for every investor), like a sudden
change in profitability of a given company; the resignation of the
board; or a change in country risk. The other extreme is private
information relevant only to an individual investor, like the
expiration of a stock-covered loan or a private analysis made. Price
Figure 2.1: Inference Problem from Price Changes
Price changes due to:

Asymmetric/differntial information Symmetric information
Information of Information of Information of

common interest common and Private interest
private interest
Decompose
Informational Noninformational
trading noise trading
Inference problem for

uninformed traders
Source: Brunnermeier (2001:28) Figure 1.1.

change may ensue from information that is both general and private,
like positive macroeconomic news modifying an investor’s risk
preference and causing him to withdraw his bank deposits to
increase the proportion of shares in his portfolio. Potential causes
of price change are summarized in the upper half of Figure 2.1.
If we classify price changes on the basis of investor decisions, we

may distinguish between two types of trading (see Figure 2.1). By
information trading we mean investor decisions based on relevant
information regarding the fundamental value of a share. Here
underlying information can be of general relevance or of both general
and private relevance. If quotation shifts because an investor took a
decision based on partly or fully private information, it may signal
“information-less” or “noise” trading.
There is a problem, however, when other investors try to deduce

the reasons behind a price change because in the case of asymmetrical
or differently interpreted information, there is no way to deduce
whether it is information or noise trading. This so-called inference
problem is present even when there is an agreement about the
impact of relevant information, i.e., the fundamental value of the
share can be given. If we add to our analysis that in most cases a
fundamental value of a share cannot be calculated (see Chapter 1),
we cannot substantiate the effect of an additional piece of news on
share price and we are faced with the deduction problem. This
holds not only for asymmetrical but for symmetrical information
as well. Noises make it impossible to correctly interpret the
correlation between stock market news and share price changes as
there is no method by which we can tell if that change was caused
by information or noise.
Share price fluctuation or volatility is also influenced by noises.

When noise trading heats up as the information content of prices
decreases, volatility might intensify. Groundbreaking studies by

Shiller (1981), and Grossmann and Shiller (1981) proved that
share price volatility is too high compared to dividend variability.
A change in volatility is closely connected to pricing uncertainties.
This effect was showed by a study by Eilifsen, Knivsflä and Saettern
(2001), which confirmed with regard to several markets that share
volatility is higher before a general assembly resolution about
dividends than after. This may be explained as more public
information lowers information asymmetry (i.e., uncertainty)
displayed in a simultaneously decreasing volatility (see also
Komáromi 2002b). The same conclusion can be arrived at from
the fact that in the non-trading hours of the day, price fluctuations
is far lower than during trading hours. During the latter hours,
noise is produced by trading itself as it means more private
information (French and Roll 1986).
Our previous analysis based on mathematical and empirical

economics showed that the formation of an asset price bubble is
strongly related to information supply – the less information
investors have (about the number of participants, timing of dividend
payment), the more likely an asset price bubble forms. Practically
the same was shown by the literary economics analyses, whereby
bubble formation was tied to a given macroeconomic shock.
Economic circumstances change and new and more efficient
production factors make share evaluation more uncertain (it becomes
more difficult to estimate future cash flow) because lacking relevant
experience and lower quality information is at the disposal of
investors. In such cases a logical and rational reaction of investors
would be not to trade at all (Black 1986). That would mitigate
pricing uncertainties, prices would diverge less from fundamentals,
and bubbles would not form. Still, if investors are to make a profit,
they have to make decisions in spite of information scarcity in a
noisy environment.
Stock market crash may be induced by the dominance of noise

trading, the cause of which is pricing uncertainty and a change in
investor behavior. In the following, we will provide a possible
psychological explanation for noise trading and make an attempt
at measuring the change in investor behavior.
2.1.1 Illusion of Knowledge in Investment Decisions

Odean (1999) showed that investors are trading in excess of the
information they receive. This happens because investors are overly
trusting the accuracy of their information and have a tendency to
overestimate their own capacities. This overconfidence fuels the
illusion of knowledge. If an investor believes decision-making
accuracy is directly proportional to the volume of information, he
is trapped in this illusion of knowledge, because even in new pieces
of information, investors are seeking out and emphasizing the
elements confirming their earlier views.
My earlier study (Komáromi 2002b) showed that not only the

absolute but also the relative amount of information is a factor in
investor behavior in a simulated market. Those surveyed not only
found a more attractive investment in a share with more information
supplied, but shareholding increased in general when the length of
the share descriptions was different. A market with such an
informational structure is more “exciting” thereby more attractive to
investors, because it creates the feeling of “something is in the air”.
Thus, irrelevant information, which is in no way related to share
evaluation, does affect the behavior of economic actors feeding on an
illusion of knowledge. This illusion is further enhanced by decreasing
transactional costs (Barber and Odean 2001; Komáromi 2002b).
An example of this is when investors have the opportunity to borrow
for purchasing shares cost-per-share goes down due to leverage.
An economic situation or change of production factors causing

a more uncertain pricing does not result in less, but actually more
trading. Noise trading not substantiated by relevant information
heightens uncertainty and investors are yet again faced with the
deduction problem. If a share price goes up, this might be
interpreted as a favorable signal and vice versa. This interpretation
is usually confirmed by post-event explanations published in the
media and hindsight bias, a psychological trait characteristic of
decision evaluation. If an investor begins to buy following a rise
and sell following a decline, he is pursuing a positive feedback
strategy. The studies by Shleifer and Summers (1990), Shiller (2000)
and Shleifer (2000) confirm that this strategy is popular among
investors and its impact is observable in price evolvement. Trend-
following behavior from investors applying a positive feedback
strategy (equivalent to a herding effect, as formally described by
Banerjee (1992)) will, as a result of the deduction problem, generate
further noise, i.e., the intensification of noise trading.
The formation of Kindleberger bubbles can thus be amended

with the observation that a rise in share prices may lower information
content and increase noise trading, which is explained by the
illusion of knowledge.
2.1.2 Co-movement of Stock Prices

Traditional pricing models take the effect of noise trading into
account only in the rise of volatility, because noise may be equally
likely to positively or negatively affect prices, i.e., investor behavior
(Komáromi 2002a). However, noise is systematic (Barber, Odean
and Zhu 2003) and may be amplified unidirectionally. Although
investors usually have heterogeneous opinions and preferences, in
a given situation, their opinions may begin to converge due to
herding effect and positive feedback.
Can we objectively measure the critical point in the opinion

change of market players, the one-sided distortion of their
expectations? Robert Shiller prepares the bubble expectations index
terminally beginning in 1989 quantifying the expectations of
institutional investors using a questionnaire. This index,
summarizing short- and long-term investor expectations, did not
show a trend in the examined period and its evolution offered no
clue as to the future movement of share prices. Shiller’s (2000)
work provides data culled from other surveys but these are not
comparable and not suitable for making general conclusions. There
is, however, a method to indirectly track investor opinions by an
examination of share price co-movement, and this shows a one-
sided amplification of noise.
Share price co-movement may have several reasons in a stock

market. The study of Morck, Yeung and Yu (2000) traced the
annual synchronous movement of share prices back to structural
and institutional factors in an international context. Structural or
fundamental factors that can be considered causes for co-movement
are ranging from the number of shares listed in the exchange to
country size, similarity of business activity, change in profitability,
etc. The details of the study offer explanation for a major proportion
of co-movement with the help of corruption indices and the legal
protection of shareholders. Their main focus was why share price
co-movement is stronger in developing countries with lower GDP
per capita than in more developed ones. Fundamental and
institutional factors (like market size, number of listed shares,
corruption index, and minority shareholder protection) all showed
that the degree of capital market maturity can mostly explain such
co-movement. If a stock market is less developed and segmented,
characterized by less institutional investors (investment and hedge
funds) and less corporate transparency by Western standards, then
share price is less of a reliable indication for investors. In such

countries, we believe the missing factors can be found in the behavior
of international investors, their portfolio reallocation principles and
rules of thumb. It is because, in spite of the above mentioned features
and the increasing uncertainty of pricing, international and domestic
investors still have to make decisions, which will inevitably result
in the amplification of noise trading.
A study by Barberis, Shleifer and Wungler (2003) showed that

the co-movement of the US stocks and the S&P 500 did not depend
on corporate news related to the shares but on investor trading
habits and, reactions to non-fundamental news. Their empirical
research showed that the price of shares would more closely follow
the path of the market index after being included in the S&P 500.
On exclusion, just the opposite was observable. According to their
explanation, investor attitude changed with regards to the particular
stocks without any underlying fundamental news. In line with the
thinking of this book, we can state that the degree of noise trading
increased with the inclusion in the index and decreased on getting
out of the S&P 500.
Daily co-movement of share prices has a different background

from synchronicity explored in the literature. While the latter can
be traced back to fundamental and institutional factors, daily share
price co-movement is dominated by trading patterns and
psychological motives on part of investors. Next, we will examine
co-movement changes over time which may be an indicator of future
investor behavior regarding shares that belong to a category of sorts.
In order to implement this, we are going to use the mathematical
basis laid down in the above-mentioned studies.
To measure co-movement, we will use the average determination

co-efficient (R2) of linear regressive equations determined by the
daily market index yield and the daily yield of a specified number
of shares:
cov t (ri ; rm )
Rt2 = , (2.1)
vart (ri )vart (rm )
where ri is the daily yield of share i; rm is the daily yield of the

market index; cov(·) is the covariance of variables; and var(•) is the
variance of the variable in the t period. The co-movement index
gained from Equation (2.1) is divided into non-overlapping equal
periods (e.g. one month) and calculate it for several periods
(t=1, 2, 3,…). The time series received this way shows the strength
of share co-movement and how it changes from period to period.
The Budapest Stock Exchange is a less developed stock market

on many parameters. In many aspects, shares listed here can be
considered as belonging to one category which is a prerequisite for
characterizing share co-movement with the indicator given in
Equation (2.1). Most of the trading in the Hungarian stock market
is carried out by foreign investors who are not able to diversify
domestically due to the small number of shares listed. While
Hungarian shares carry a market risk of 85-85 percent and an
individual risk of 5-15 percent, in the case of the New York Stock
Exchange this proportion is just the opposite. Due to the low activity
of domestic institutional and private investors, prices are determined
by the portfolio decisions of foreign investors, which domestic
investors may interpret as signals, probably integrating them into
their strategies (Pound and Shiller 1986).
Figure 2.2 shows the co-movement index of the 13 highest volume

shares listed in the BSE from January 1996 to June 2003. To calculate
the determination co-efficient, one-month period was chosen and
BUX, the official index of BSE, was used as market index. The average
value of the share co-movement index was significantly higher in the
Figure 2.2: Daily Share Co-Movement in the

Budapest Stock Exchange
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Jan-96
May-
Sep-96
Jan-97
May-
Sep-97
Jan-98
May-
Sep-98
Jan-99
May-
Sep-99
Jan-00
May-
Sep-00
Jan-01
May-
Sep-01
Jan-02
May-
Sep-02
Jan-03
May-
Note: The thin line refers to the share co-movement index, calculated for the
following stocks: Matáv, MOL, OTP, Borsodchem, Egis, Inter-Európa Bank, Pick,
Pannonplast, Prímagáz, Zwack, Richter, Rába, TVK, Fotex. The thick line refers to
the 5-member moving average of index values.
Source: own calculation based on www.portfolio.hu quotation data.
period from September 1997 to April 1999 than in other periods,

meaning a high variability in co-movement, the increase of which
implies a shift towards noise trading.
We believe the co-movement index calculated above for the

Hungarian market is a suitable tool to track noise trading because
BSE is characterized by a small number of shares, dominance of
foreign investors and little room for diversification. In such cases
co-movement index may be a signal of large scale price decline
later. The Hungarian example will be explored in a wider context
in Chapter 3. For more advanced capital markets, it is necessary to
refine co-movement measurement and to create sophisticated
indices, which can be an objective of further research.
The price co-movement of stocks, other financial assets and

investments not justified by fundamental reasons is a recurring theme
of historical bubbles discussed in Chapter 1.3. Kindleberger (2000),

Galbraith (1994) and Shiller (2000) mention several examples where
optimistic investors could not distinguish in the course of a stock
market boom between realistic enterprises based on well-grounded
ideas and companies founded on insecure basis or fraud. As a
consequence, all shares of the hyped industry of the era were rising
almost continuously. An example of this is Internet- related firms
during the dotcom boom at the end of the 1990s and the
uninterrupted success of serial capital issues by those dotcom firms.
Henceforth, we will make use of share price co-movement not

in a mathematical but a literary analysis, accepting the hypothetical
assumption that price co-movement not justified by fundamentals
can signal non-information based, i.e., noise trading, forecasting
that may result in a future price collapse and a crash.
2.2 How to Distinguish Stock Market Bubbles?

In the following, we will list typical features of bubbles using the
elements of the Kindleberger definition, outlining stock market
bubble anatomy.
The formation of a bubble starts with the clear and continuous

rise of share prices caused by an exogenous shock affecting the
economy. This initial displacement influences future outlook in a
positive way, generating expectations of further rise. If share prices
distinctly begin to rise, uninformed investors, partly due to the
deduction problem, take this as a positive signal. Share of particular
industries and companies may become popular. New buyers appear
in the market and the proportion of shares increases within
portfolios causing a surge in trading volume. As many investors are
pursuing a positive feedback strategy, this coupled with the lack of
relevant information will amplify noise trading.
A stock market boom can be described as a bubble if there is

high probability of a large scale fall in share prices. Stock market
crash is not triggered by fundamental news or by a certain level of
share overvaluation. Instead, it happens because of a drastic change
in the behavior of market players. This is why the necessary and
sufficient conditions for the bursting of a given asset price bubble,
applicable in practice, cannot be provided with the tools of
mathematical economics. A market crash will ensue with a high
likelihood if noise trading becomes dominant, the signals of which
are to be found in the following stochastic factors:
• Increasing effect of leverage. As a direct consequence, more
money is at the disposal of investors (see previous paragraph).
If investors borrow to buy shares, have the opportunity to
postpone payment, or making a purchase without full financial
cover, it is impossible for them to realize long-term profit on
that particular stock, i.e., they are unable to make dividend
payment. This means a short sale constraint shortening the
average investment period. The due date of debt repayment is
private information incurring, on the one hand, deduction
problem and noise trading. On the other, if there is an
increasing pool of leveraged shareholders, repayment date and
a short sale constraint will more likely be due at a given
moment, amplifying the degree of the price fall.
• Increasing activity on part of the economic policy. Economic
policy, and monetary policy in particular, can directly
influence the conditions of credit, bond and money markets
connected to stock markets, thus making the state a
protagonist in the stock market. Intended monetary
expansion or restriction is always a signal, as it attempts to
stimulate or curb the rise of prices. For example, the frequent
and tendentious revisions of the base rate convey a series of

signals towards market players. In theory, the opportunity
cost of shares (the rise in bond yields) prompts investors to
lower the share of stocks in their portfolios. Sometimes,
however, investors are late and inaccurate in integrating
signals of the economic policy into their expectations,
increasing the volume of noise in the market.
• Increasing number of corporate scandals, fraud and corruption.
Share price rise augments the power and influence of executives,
while directly affecting their wealth through managerial stock
options. Information asymmetry enables them to use methods
verging on fraud to maintain the trust of owners-shareholders
if corporate performance is not contributing positively to the
share price. The disclosure of such cases may undermine trust,
causing a change in investor behavior and prompting the sales
of the shares of other companies.
• Fundamentally unjustifiable co-movement of share prices.
The co-movement of different shares or investments may
signal a dominance of noise trading. When investors do not
evaluate a given asset based on its expected future yield, i.e.,
do not evaluate an enterprise based on the probability of its
future success, and instead they make simplifications and
use rules of thumb, a fundamentally unjustifiable share price
co-movement may ensue. If this co-movement increases, price
fluctuation may signal a dominance of noise trading,
forecasting a stock market collapse.
The last characteristic of stock market bubbles is that the boom
and subsequent crash must have an impact on the economy. Only
then will the natural instability of stock markets become a factor
affecting economy, without which the concept of a bubble would
be weightless. By negative impact we mean a slowdown in economic

growth or a decline in consumption and/or investment. However,
a bubble may carry positive impacts as well which display themselves
either during the boom or following the crash, in the long run.
One such effect is the facilitation of capital issue for a given industry
allowing a better financing of riskier solutions and developments.
After a crash, the framework surrounding the stock market may
also change, bringing about legal, regulatory and institutional
evolution as a consequence of the collapse. If a stock market boom
has no impact on the economy of a country or on related regulation
and institutional structure, we contest such a phenomenon can be
called a bubble.
Initial displacement, distinct price rise, new buyers (increasing

trade volume) all are direct traits of a bubble, while leverage, the
large number of economic policy signals, corporate scandals, fraud
and corruption are indirect indicators of the phenomenon. In light
of the above description, we sum up in Table 2.1 the most important
and the most recent stock market bubbles in economic history,
using the studies by Kindleberger (2000) and Shleifer (2000).
Table 2.1: (I.) Characteristics of Stock Market Bubbles in

Economic History
Period 1710-1720 1717-1720
Country England France
Speculation on Stocks of South Sea Stocks of Mississippi
Company. Company.
Initial displacement New markets, trade New market, trade with
with South America. Louisiana.
Undertaking and Conversion and
conversion of securitization debts.
Government debts.
Fueling investors positive Easier public offerings in Easier public offerings in
attitude succession. succession.
Support of Government Laws full powers in
(Bubble-Act). financing decisions,
and absolute support of
Government.
Signals for the danger Insider trading involving Leverage Critiques on
of crash MPs, Leverage. unsophisticated
investors (insider
information);
Administrative steps.
Crash August 1720 May December 1720
Macroeconomic / Bubble-Act made it Reform of finance of
regulatory effects difficult to found new State started and came
joint-stock companies. to a sudden standsill
after the crash.
Period 1845-1847 1873

Country England USA
Speculation on Stocks of railway Stocks of railway
companies companies
Initial displacement New, fast and cost Expansion after
efficient way of transport American civil war
Contd...
Contd...
Fueling investors Development of Large financial
positive attitude infrastructure by government support for
Government. railways.
Signals for the danger Insider trading (connection Leverage (short term
of crash between London Society loans from Europe).
and George Hudson).
Leverage (dividends are
financed by public
offerings).
Crash October 1847 September 1873
Macroeconomic / Suspension of Banking Because of development
regulatory effects Act in 1844. of infrastructure,
Reform in Accounting economic activity
(dividends can be paid increases in western
from profits). states and trade with
eastern states.
Table 2.1: (II.) Characteristics of Stock Market Bubbles in

Economic History
Period End of 1880s 1920s
Country England USA
Speculation on Stocks of companies Stocks after IPOs
interested in agricultural
lands in Argentina.
Initial displacement Increase in demand of End of deflation fears
Argentine agricultural after WWI.
products. Increase in mass
production.
Fueling investors Increase of number of Development of
positive attitude public companies. financial services
Political incentive of (investment fund).
Argentine state Easier public offerings in
president. succession.
Contd...
Contd...
Signals for the danger Biased analyses (Barings), Leverage, buying shares
of crash insider trading. on margin.
Actions of politics and
economic policy.
Crash November 1890 October 1929
Macroeconomic / Coup detat in Argentina. Glass-Steagall Act,
regulatory effects Restriction of foreign Chinese wall between
investments. (separating) banking and
investment services
(SEC).
Splitting big companies.
Deepening economic
crisis.
Period 1982-1987 1990s

Country USA South-East Asian
Countries*
Speculation on Stocks of reorganized Stocks, Government
companies bonds
Initial displacement Waves of merger & Increase of speed of
acquisition because of economic growth, cost
cost efficiency efficiency of
reorganizations.
Deregulation,
liberalization.
Fueling investors Government incentives Fix exchange rate
positive attitude for economic system
development. Government incentives
Development of financial for economic
services, tax-cuts. development (tax
benefits).
Signals for the danger Junk-bonds. Manipulating financial
of crash Insider trading reports.
(management). State budget deficit
Contd...
Contd...
Program-trading. and expansive economic
policy.
Crash October 1987. December 1997, January
1998.
Macroeconomic / Changes in trading rules. Domino-effects in
regulatory effects Brady-Commission. emerging security
Making buying outs more market.
difficult, less tax benefits. Restrictions of capital
flows.
Table 2.1: (III.) Characteristics of Stock Market Bubbles in

Economic History
Period 1990s 1990s
Country Russia* Brazil*
Speculation on Stocks and government Stocks and government
bonds. bonds.
Initial displacement Positive signs of Positive consequences
economic recovery after of stabilization pack of
collapse of USSR 1993-94.
deregulation, Deregulation,
liberalization. liberalization.
Fueling investors positive Fix exchange rate system. Fix exchange rate
attitude Government incentives for system.
economic development. Government incentives
for economic
development.
Signals for the danger Large state budget deficit. State budget deficit and
of crash Corruption and scandals in expansive economic
business and state sectors policy.
(e.g., privatization). Frauds in financial
reporting.
Crash August 1998 January 1999
Macroeconomic / Domino-effect to Eastern Domino-effect to Latin
regulatory effects and Central European America.
countries.
Contd...
Contd...
Period 1995-2001
Country USA and developed countries
Speculation on Dotcom companies.
Initial displacement Computerization (PC) in a wide range:
Development of Internet households, firms.
Fueling investors positive New channels of communication.
attitude Cost-efficient and fast trading, marketing, etc.
Signals for the danger Activity of monetary policy.
of crash Frauds and scandals in big firms.
Crash March-April 2000
Macroeconomic / After economic recession, robust economic
regulatory effects growth.
Debates of corporate governance in a wide
range.
Regulation on more transparent investment
decisions.
* Note: Stock market bubbles in Eastern and South Asia, Russia and Brazil were
accompanying events of currency crises. In these countries, stock markets do not play
a great rule like in Anglo-Saxon countries, but the prices of certain shares and market
indexes drive investors’ attention. There were domino effects in related regions and
contagion in other securities. These effects had negative impacts on the FDI to the
regions, and worsened investors’ appreciation.
This table consists some quotations from Galbraith (1994) Kindleberger (2000),
Shiller (2000) and Shleifer (2000).
!
The Hungarian Case, 1996-2003
“A change of political system takes 7 month, an economic one at

least 7 years. A change in people’s attitude may even take 70 years.”
Árpád Göncz,
(the President of Hungary 1990-2000).
* SE re-opened in 1990 after a 42-year break as one of the crucial

institutions of post-socialist market economy. Of East-Central
European countries, Hungarian economy was the most open, and
as a consequence, domestic securities market began functioning a
decade earlier (Rotyis 2001:167). The first, and for some time the
only listed company was IBUSZ, which saw its share price multiply
in the course of a couple of month and then falling back to its
initial issue level. The private sector was active in creating financial
institutions: during the first years after BSE resumed, many private
companies tried to issue capital through the stock market. At the
beginning of the decade, a large number of domestic equity traders
and brokerage firms were actively participating in the market as
well as an increasing number of foreign companies.
Still, until 1996 we cannot speak of an active domestic stock

market because of the low number and low liquidity characterizing
The Hungarian Case, 1996-2003 89
the Hungarian exchange. The period between 1996 and 2003 was
more eventful both for BSE and investors, thanks to the privatization
and public listing of previously state-owned companies. In the stock
market, foreign investors encountered a full liberalization which
was soon complemented by the liberalization of the market for
derivatives and Hungarian government securities embodying short
– and long-term debt. Rotyis (2001) believes that these were the
market reasons why the Hungarian securities market became more
internationalized than neighboring markets, more closely
connecting to foreign stock exchanges.
In the following, we will examine whether the boom and bust

experienced by Hungarian shares can be considered a bubble with
mostly domestic cause or mostly domestic economic impact.
Figure 3.1 shows the change of BUX, the official index of the
BSE between 1996 and 2003. Three distinct periods can be
Figure 3.1: BUX Between January 1996 and September 2003

11000%
10000%
9000%
8000%
7000%
6000%
5000%
4000%
3000%
2000%
1000%
0%
Dec-95
Jun-96
Dec-96
Jun-97
Dec-97
Jun-98
Dec-98
Jun-99
Dec-99
Jun-00
Dec-00
Jun-01
Dec-01
Jun-02
Dec-02
Jun-03
Source: www.portfolio.hu
identified during the evolution of BUX. During the first boom,

BUX rose from 1,500 points in January 1996 to 9,000 points by
the spring of 1998. This was followed by a fast, 50 percent dive

lasting until September 1998. The second boom reached its high
in the spring of 2000 with 10,472 points. The third period saw
BUX fluctuate in the 6,000-9,000 range. We will now identify
the reasons, that are directly related to the Hungarian exchange,
behind the rise and fall of BUX in light of the bubble parameters
discussed in the previous chapter.
The shock causing the initial displacement was provided by the

1995 Bokros-package and the stock exchange privatization of state
companies. Economic stabilization measures lent credibility to
economic policy, while crawling peg lowered exchange rate risk of
forint. These were prerequisites for the involvement of foreign
investors in the stock market privatization of Hungarian firms.
Between 1995 and 2000, approximately 40 percent of capital
inflow arriving in Hungary was related to companies listed on BSE.
Not only foreign, but domestic institutional and individual investors
Figure 3.2: Trade Volume and Turnover Velocity Between
1996 and 2003
140%
120%
100%
80%
60%
40%
20%
0%
1996 1997 1998 1999 2000 2001 2002 2003
Note: Dotted line = trade volume; 1998 = 100%.

Solid line = turnover velocity (annual volume divided by annual average
capitalization).
Source: own calculations based on data from www.bse.hu and www.fese.org.
also had the opportunity to buy shares in companies like TVK,

BorsodChem, OTP, MOL, Pick.
In the middle of the decade BSE trading volume grew at an

enormous rate in absolute value. The number of listed shares
continuously grew until 1998 and reached 50. Trade volume went
up even compared to market capitalization, i.e., turnover velocity
increased (see Figure 3.2). After 2001 investor activity declined so
we definitely cannot speak of a bubble in the third period.
Internationally BSE is a small stock exchange and only a segment
within the Eastern-Central European market in the eyes of foreign
investors. Trade volume rise was largely due to such foreign investors,
pulling domestic investors with them until 1998. The first fall in
share prices in 1998 was affecting domestic investor more, which
is shown by the fact that they owned 30 percent of BSE
capitalization in 1997 and less than 20 percent by the end of 1999
(see Figure 3.3).
Figure 3.3: Structure of Share Ownership in BSE Between

1997 and 2003
80%
70%
60%
50%
40% 12
12 12 12
12 12
1212 12 11 11 11 11 12
30% 12
12121212
12 12 1212 1212
1212
1212 11 1212
1212 1212
1212 1212
1212
12 11
20% 12 1212 12 11 11 11 11 11 12 12 12 12 12 12 11
10% 1212121212121212
12 12 12 1212 12 1212
12 11
11
11 11 11 11
121212
12 12 1212121212
12 12 12
12 11
12
1212
1212
1212
1212
1212
1212
1212
1212
12 1 11 1 1 1 1 12
1212
1212
1212
1212
1212
1212
12 1
0%
Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2
1997 1998 1998 1999 1999 2000 2000 2001 2001 2002 2002 2003
Note: The black, grey and white bars show the proportion of foreign, domestic
and individual domestic investor ownership, respectively.
Source: www.mnb.hu
Similarly to the first period, between spring 2000 and autumn

2001, BUX dropped by almost 50 percent. This decline, however,
took place during a longer period. The change in shareholder
structure (Figure 3.3) shows that the price change of the millennium
was in line with a gradual decline in the proportion of foreign
investors, which means we can speak of a crash only in the first
case, if at all.
Domestic investors suffered heavier losses, not only in prompt

trading but also in the ever-more significant BUX futures contracts.
Another form of leverage was the loan-based stock buying schemes
advertised by securities trading firms, which became very popular
with individual investors by 1998. The daily co-momement of share
prices indicates that noise trading grew dominant (see Figure 3.4),
which suggests that the crash of 1998 could have happened because
Figure 3.4: Daily Co-Movement of BUX and Shares
Between 1997 and 2003
120%
100%
80%
60%
40%
20%
0%
Jan-96
May-
Sep-96
Jan-97
May-
Sep-97
Jan-98
May-
Sep-98
Jan-99
May-
Sep-99
Jan-00
May-
Sep-00
Jan-01
May-
Sep-01
Jan-02
May-
Sep-02
Jan-03
May-
Note: Dotted line = BUX (maximum = 100%), Thin line = stock co-momement
index, which is the average R2 of linear regression equations of market and
individual share daily returns (Matáv, MOL, OTP, BorsodChem, Egis, Inter-Európa
Bank, Pick, Pannonplast, Prímagáz, Zwack, Richter, TVK, Fotex). Thick line =
5-member moving average of index values.
Source: Own calculation based on price data from www.portfolio.hu.
of a drastic change in investor behavior. A reasons for this could

include portfolio reshuffling on part of foreign investors following
the Russian crisis and the defection of individual investors.
The end of 1999, start of 2000 boom and bust already happened
at lower investor activity and trading volume. BUX evolution was
by then affected only by foreign events, like the burst of the US
Internet bubble. BSE trade volume fell dramatically in the second
half of 2000 and in 2001 resulting in the further marginalization
of the Hungarian exchange.
As an important feature of a possible Hungarian stock market

bubble, we have to take note of macroeconomic impact. Examining
several factors, however, it seems that in the past period, the Hungarian
exchange could not have much macroeconomic impact. Following
the 1996 boom characterized by the privatization of state-owned
companies, no major public offering happened. After this BSE volume
dropped to a fraction of its previous highs, proceeds from privatization,
constituting a major part of FDI inflow were probably not dependent
on the existence of the stock market even in the first period. This
confirms implicitly that the Hungarian exchange did not offer a
risk-spreading opportunity for foreign investors because domestic
investors’ contribution to market liquidity was limited. We can find
no other signs, either, suggesting Hungarian economic policymaking
was influenced by stock market events.
Companies did not make use of the capital issue opportunities

offered by the stock exchange, this being both a cause and an effect
of the small economic weight of BSE. The new Securities Act
ordering the listing of companies based on some criteria did not
bring the intended results, which is again a proof that economic
problems often cannot be tackled by legal methods. An attempt
on the part of BSE to lower listing costs was not an adequate
motivation for companies. Behind a preference for bond or debt

market lay a non-Anglo-Saxon corporate governance style and
conscious management decisions. The latter were aimed at the
preservation of independence which is only compatible with public
offerings if a large number of individual investors can be reached.
In September 1998, stock proportion in household assets was

almost 9 percent, gradually declining afterwards. Households keep
only a marginal proportion (13 percent), of their assets in securities,
and only 2 percent in stocks, in particular. Another 9 percent is in
investment funds but only a fraction of this (3 percent) is held in
Hungarian shares. This drastic fall in the proportion of shares did
not have any significant impact on consumption, because it was
this very period in which the gradual indebtedness of Hungarian
households began.
The low number of domestic investors, and particularly

individual investors, is the main reason behind the lack of continuous
liquidity, a staple of a well-functioning market in the Hungarian
exchange. Today BSE trade volume sometimes reaches 2000 levels
but 90-95 percent of this volume is still attributed to 4-5 shares.
There are no new offerings in the Hungarian market but perhaps
there would not be much interest in them as the outdated savings
structure of Hungarian households shows no evolution.
Underdeveloped Hungarian financial culture is the reason why the
economic importance of BSE is still marginal. Based on the approach
of this book, we argue that the booms and busts experienced in
BSE between 1996 and 2003 cannot be termed bubbles.
Summary 95
"
Summary
6 his book is aimed at providing a definition of “stock market

bubble” filtering out the contradictions and weaknesses of
relevant theories. First, we presented existing definitions and
unresolved contradictions using the tools of mathematical and
literary economics. An example of this was the issue of fundamental
value and excessive speculation. We summed up investor behavior
characteristics diverging from rationality, the diverse results of
laboratory experimental economics, and the features of the most
important stock market bubbles in economic history.
A typical bubble formation was described as follows:
The emergence of a stock market bubble starts with a robust

and continuous price rise usually caused by an exogenous
macroeconomic shock. This initial displacement affects expectations
regarding the future in a positive manner. Stock market trade volume
soars in line with an increase in noise trading. A boom like this can
be called a bubble if the probability of a drastic decline, i.e., crash,
grows. Only a real macroeconomic or regulatory impact of the crash
can lend economic weight to the term “bubble”.
An increasing likelihood of crash is, therefore, indicated by the

intensification of noise trading. Signs of such intensification include
an increase in the activity of economic policymakers; news related
to corporate scandals, frauds and corruption; leverage; and in some
cases share co-movement. The final trait of a bubble is the most
important one, because this emphasizes the basic wealth reallocation
role of stock markets. Stock market instability will inevitably incur
boom-bust cycles, of which actual bubbles can be identified using
the above criteria, making it a concept with economic relevance.
Stock market bubbles do exist because the term is used by

economists to describe certain market phenomena. This book
provides a possible answer to what phenomena we can call bubbles
i.e., what indicators suggest as a “qualified case” of stock market
boom. This answer is only one of many, but we believe it is more
accurate and valid than previous ones.
The examination and analysis, in light of this definition, of

additional international and domestic cases can be the topic of
further research, as well as the assembly of a comprehensive set of
causes behind share co-movement, sorting out methodological
problems; and preparing relevant comparative studies.
Summarizing the Messages of this Book in Theses

Theses* (1) – (5) summarize and recompose the theoretical,
and contradictory results of related literature, which are the basic
points of the book. The purpose of my work is to clear these
contradictions and to put theoretical problems into a new
framework, and finally to give a more appropriate explanations for
stock market bubbles. Further propositions comprise the essence
of new results and conclusions of the book. Proposition (6) sheds a
* Based on the theses of my dissertation (Komáromi 2004).
Summary 97
light on the relationship between investors’ behavior and a stock

market bubble, and explains the role of noise trading. Points (7) –
(9) summarize the phenomena that may be signals for the
dominance of noise trading. Thesis (10) is the main statement of
the book, it gives typical features of stock market bubbles. In
proposition (11) we express the effects of stock market bubbles on
real economy. Propositions (11) – (12) explain the experience of
two cases: Dutch Tulip Mania and Hungarian booms and crashes
between 1996 and 2003. Each proposition contains references to
the chapters of the book in brackets.
(1) Mathematical economics defines asset price bubble as a

positive difference between actual and fair (fundamental)
prices of the asset. On the contrary, “bubble” in verbal
(literacy) economics usually covers (i) a more general,
broader economic phenomenon, when asset prices increase
significantly and continuously, which is fueled by investors’
expectation for further increase, and (ii) it may be
accompanied by macroeconomic effects (Kindleberger’s
definition). The two definitions seem to be different.
However, they basically mean the same: stock prices
definitely deviate from economic fundamentals. The
difference stems from the different sets of tools used by
two approaches (mathematical and verbal economics).
[1. Introduction; 1.1 Introduction; 1.3 Introduction]
(2) The strong and irrational deviation of a stock or stocks
from fundamental value (overvaluation) cannot be proved.
One of its reasons is that it is difficult to give the
fundamental value of a financial asset and its change in
practice. If there is no doubt that a stock price obviously
differs from its fundamental value (e.g., closed-end funds,
twin-stocks), we cannot decide whether we face with

over – or undervaluation. [1.1.1; 1.1.2; 1.1.3; 1.1.5]
(3) In the laboratory experimental simulations, we can directly
observe how bubbles occur most of the time, because future
dividends are previously given, and selling/buying decisions
determine stock prices. On the other hand, in these
experiments we cannot filter out the distorting effects of
laboratory environment. Even then, when changing the
conditions in experiments, we can conclude that the extent
of price bubbles depends on the liquidity and the
information subjects have. [1.2]
(4) Investors essentially buy a stock to obtain its future returns
(dividends, profits from selling at higher prices). Apart
from frauds and swindles, any future dividends can be
expected that may justify the actual stock price. In the
case of stocks, we rarely face with Ponzi-financing. On the
contrary when we have loans, for instance, paying interests
can only be financed by other new loans. [1.1.5]
(5) Kindeberger’s definition is not appropriate to differentiate
regular fluctuations of stock prices from economic and
scientific aspect. The weakness of his argument comes from
not giving the standard level of speculation. On the other
hand, there is no sense in defining such a level; we have to
define bubbles in a different way. [1.3 Intr.; 1.3.5]
(6) I strongly argue that a stock market bubble should be
defined as a consequence of investors’ behavior. I disregard
other possible economic reasons; their roles are not
mentioned in my definition, because in stock markets,
prices primarily reflect investors’ expectations (see Keynes).
In fact, investors just take bets on future prospects of listed
firms. Expansions and bursts of bubbles can be traced back
Summary 99
to specific features of investors’ behavior, especially

overconfidence. “Overconfidence” implies that investors
would be better off not trading on the available information,
but nevertheless they trade. When it happens, investors
have an illusion of knowledge, which is accompanied by
increase in public and private information, and irrelevant
information (noise). To put it other way, when a stock
market bubble occurs, the intensity of noise trading
increases too. [1.1.3; 2.Introduction; 2.1; 2.1.1]
(7) When it is said the economic policymakers become more
active, it implies an increase in the noise trading. Different
policy actions are signals for the investors, and drive
investors’ attention, and investors take these signals into
account in their expectations but with delay and
inaccuracies. Same effect can be found when number of
news about scandals, frauds, and corruption increases, and
these signals may indicate sales of stocks and not of related
firms. [1.3; 2.2]
(8) Leveraged trading is noise trading in one respect, because
finite duration of the particular asset involves short sale
constraint, and investors do not trade on public but private
information and in consequence of private constraint
(deadline of repayment of private loan). Leverage also
increases liquidity of investors involved. [1.3; 2.2]
(9) If stock prices move together, and no fundamental factors
justify this synchronicity, it may indicate noise trading.
There is no standard level of price co-movement, but if it
rises significantly without change of any economic or
market factors, it may show that investors’ decision-making
progress is becoming unsophisticated. To measure price
synchronicity we can take average R-squared between stocks

and market index, if the stocks belong to one industry or
one well-defined market. In these cases, investors regard
the stocks as one bunch, and the increase of co-movement
rises to the level of noise trading. [1.3; 2.1.2; 2.2]
(10) We distinguish bubbles from other fluctuations of stock
market with the following feature: (i) A stock market bubble
starts with a strong and continuously rising stock prices,
mostly due to a macroeconomic shock (ii) This initial
displacement positively affects investors’ expectations on
the future. The volume of stock market also rises
significantly, as the noise trading increases. We regard booms
as bubbles if the probability of large price drop – market
crash – is considerable. (iii) Finally and probably the most
important feature of stock market bubbles is the real effect
at macroeconomic or regulating level. We distinguish
bubbles from regular fluctuations caused by instability of
stock markets with the said features. Consequently these
characteristics give the economic importance to the term
“bubble”. [2.2]
(11) Trading in stock exchanges is basically a zero-sum game;
its role is only to distribute wealth, but indirectly has effects
on macro- and micro-economic levels. This is a
distinguishing feature of bubbles. There are some negative
effects, when stock prices are rising, and firms make over-
and mal-investments financed by public offerings. These
decisions have repercussions on firms’ revenues and cash
flows, they also increase investors’ risk. Other well-known
effect of stock market booms is the wealth-effect. If value
of household-owned stocks increases, their consumption
Summary 101
also rises, and it may follow from the foregoing that inflation
may accelerate. At micro level, firms may easily obtain quasi-
venture capital when market is soaring. In these periods
investors make decisions on less information or noise. It
may give an impulse to the industry and the economy as
well. Another positive output is when market crash forces
important changes in regulatory environment, and the
efficiency of market may improve. [1.3; 2.2]
(12) If we use the bubble-description of the book, we cannot
classify the Dutch tulip-speculation as a typical bubble.
First reason is that there are no reliable data. Apart from
missing sources, this speculation could not cause any real
effects on the economy or the regulatory environment in
Netherlands. [1.3.1]
(13) Booms at Budapest Stock Exchange (BSE) between 1996
and 2003 are not considered as bubbles. In the Hungarian
stock market, some bubble-phenomena (co-movement of
prices, leverage) can be seen between 1997 and 2000, but
the BSE plays an insignificant role for financing firms or
accumulates savings in Hungary. In this period Hungarian
firms were not active in raising their capital through public
offerings. Other signal that supports the statement above,
is that the stock exchange had not been prerequisite for
foreign capital inflow to Hungary. Some formerly state-
owned companies (e.g., MATÁV, MOL, OTP) became
privatized and stakes were sold directly to foreign investors,
but insignificant activity of domestic investors did not mean
risk-sharing opportunities for them. [3]
References
Allen, F., Morris, S. and Postlewaite, A. (1993): “Finite Bubbles with Short
Sale Constraints and Asymmetric Information”. Journal of Economic Theory
Vol. 26. pp. 201-229.
Bácskai T. (2003): “Bubbles and recession. Allocation, growth and
equilibrium”. Bank and Tõzsde Vol. 11. No. 1. p. 12. (in Hungarian).
Baker, D. (2000): “The Cost of the Stock Market Bubble”. Working paper,
Center for Economic and Policy Research, USA. Nov. 2000.
Banerjee, A. V. (1992): “A Simple Model of Herding Behavior”. Quarterly
Journal of Economics Vol. 107. No. 3. pp. 797-817.
Barabási A.-L. (2002): Linked. First Plume Printing, Penguin Group, NY,
USA, 2003.
Barber, B. M. and Odean, T. (2001): “The Internet and the Investor”.
Journal of Economic Perspectives Vol. 15. No. 1. pp. 41-54.
Barber, B. M., Odean, T. and Zhu, N. (2003): “Systematic noise”. Working
paper, University of California, Davis, USA. Apr. 2003.
Barberis, N., Shleifer, A. and Vishny, R. (1998): “A model of investor
sentiment”. Journal of Financial Economics Vol. 49. pp. 307-343.
Barberis, N., Shleifer, A. and Wurgler, J. (2003): “Co-movement”. Working
paper. Harvard University, USA.
Barberis, N. and Thaler, R. (2001): “A Survey of Behavioral Finance”.
Working paper, Harvard University, USA.
Bishop, M. (2003): “Revenge of Dotcom”. In: The World in 2004 - The
Economist and Magyar Hírlap Special issue Dec. 2003. pp. 72-73. (in
Hungarian)
Black, F. (1986): “Noise”. Journal of Finance Vol. 41. No. 3. pp. 529-543.
Blanchard, O. J. (1979): “Speculative bubbles, crashes and rational
expectations”. Economic Letters Vol. 3. pp. 387-389.
Bogan, V. (2003): “Rationality and Internet Bubble: Theory and Evidence

from IPO Markets”. Working Paper, Department of Economics, Brown
University, USA.
Bõgel Gy. (2002): “The New Economy and Schupeter” In: Vállalat,
Információ, Tudomány. Alma Mater Series, Budapest University of
Technology and Economics, 2002. pp. 141-158. (in Hungarian)
Brunnermeier, M. K. (2001): “Asset Pricing under Asymmetric
Information: Bubbles, Crashes, Technical Analysis and Herding”. Oxford
University Press, Oxford, UK.
Caginalp, G., Porter, D. and Smith, V. (2001): “Financial Bubbles: Excess
Cash, Momentum, and Incomplete Information”. Journal of Psychology and
Financial Markets Vol. 2. No. 2. pp. 80-99.
Cecchetti, S. G. (1992): “Stock market crash of October 1929”. In:
Newman, P. and al. (eds.): The New Palgrave Dictionary of Money and
Finance. III., Macmillan Reference, London, 1992. pp. 573-577.
Cont, R. and Bouchaud, J., P. (1998): “Herd behavior and aggregate
fluctuations in financial markets”. Working paper, Science and Finance,
France. Jan. 1998.
Csontos L., Király J. and László G. (1997): “The great tremble at
Millennium”. Közgazdasági Szemle Vol. 44. Nos. 7-8. pp. 569-596. (in
Hungarian).
Daniel, K., Hirshleifer, D. and Subrahmanyam, A. (1998): “Investor
Psychology and Security Market Under- and Overreactions”. Journal of
Finance Vol. 53. No. 6 pp. 1839-1885.
De Long, J. B. (1992): “Noise trading”. In: Newman, P. and al. (eds.): The
New Palgrave Dictionary of Money and Finance. II., Macmillan Reference,
London, 1992. pp. 34-36.
De Long, J. B., Shleifer, A. and Summers, L. H., Waldmann, R. H. (1990):
“Noise trader risk in financial markets”. Journal of Political Economics Vol. 98.
pp. 703-738.
Eilifsen, A., Knivsflå, K. H. and Sættem, F. (2001): “Earnings
announcements and the variability of stock returns”. Scandinavian Journal
of Management Vol. 17. pp. 187-200.
References 105
Farmer, J. D. (1999): “Physicists Attempt to Scale the Ivory Towers of

Finance”. Computing in Science and Engineering, Vol. November/December
1999. pp. 26-39.
Flood, R. P. and Hodrick, R. J. (1990): “On Testing for Speculative
Bubbles”. Journal of Economic Perspectives Vol. 4. No. 2. pp. 85-101.
French, K. R. and Roll, R. (1986): “Stock Return Variances – The Arrival of
Information and the Reaction of Traders”. Journal of Financial Economics,
Vol. 17. pp. 5-26. North Holland.
Friedman, M. and Schwartz, A. J. (1963): “A Monetary History of the
United States, 1867-1960”. Princeton University Press, Princeton, New
Jersey, USA.
Froot, K. A. and Obstfeld, M. (1989): “Intrinsic Bubbles: the Case of Stock
Prices”. Working paper w3091, NBER, Cambridge, MA, US.
Galbraith, J. K. (1994): A Short History of Financial Euphoria. Reprint,
Penguin NY.
Garber, P. M. (1990): “Famous First Bubbles”. Journal of Economic
Perspectives Vol. 4. No. 2. pp. 35-54.
Garber, P. M. (1992): “Crashes”. In: Newman, P. and al. (eds.): The New
Palgrave Dictionary of Money and Finance. I., Macmillan Reference,
London, 1992. pp. 511-513.
Garber, P. M. (2000): “Famous First Bubbles: The Fundamental of Early
Manias”. MIT Press, Cambridge, Massachusetts, London, England.
Gilles, Ch. and LeRoy, S. F. (1992): “Asset price bubbles”. In: Newman, P.
and al. (eds.): The New Palgrave Dictionary of Money and Finance. I.,
Macmillan Reference, London, 1992. pp. 74-76.
Gregory, C. A. (1992): “Mississippi Company”. In: Newman, P. and al.
(eds.): The New Palgrave Dictionary of Money and Finance. II., Macmillan
Reference, London, 1992. pp. 714-715.
Grossman, S. J. and Stiglitz, J. E. (1980): “On the impossibility of
informationally efficient markets”. American Economic Review Vol. 70. pp.
393-408.
Grossman, S. J. and Shiller, R. J. (1981): “The Determinants of the

Variability of Stock Market Prices”. American Economic Review Vol. 71. pp.
222-227.
Hicks, J. R. (1946): “Value and Capital” 2nd edition, Oxford University
Press, Oxford, UK.
Keynes, J. M. (1936): The General Theory of Employment, Interest and Money.
Macmillan Cambridge University Press. Cambridge, US.
Kindleberger, Ch. P., (2000): Maniacs, Panics, and Crashes: A History of
Financial Crises. 4th edition John Wiley and Sons, New York, USA.
Kindleberger, Ch. P. (1991): “Bubbles”. In: Eatwell, J. and al. (eds.): The
New Palgrave, The World of Economics. Macmillan Press Limited, UK. 1991.
pp. 20-22.
Komáromi Gy. (2000): “Why Behavioral Finance?” In : Doctorands’ Forum
2000. Faculty of Economics - Section Series University of Miskolc, 2000.
pp. 27-34. (in Hungarian)
Komáromi Gy. (2002a): “Theoretical and practical relevance of the efficient
market hypothesis” Közgazdasági Szemle Vol. 49. No. 5. pp. 377-395. (in
Hungary)
Komáromi Gy. (2002b): “Why have the stock markets become noisier after
the revolution of information and technology?” In: Evolutions of
Institutions and the Knowledge Economy Conference Proceedings.
University of Debrecen, 2002.
Komáromi Gy. (2003a): “How does the illusion of knowledge impact on the
stock prices ?”, In: Kovács Z. (ed.): Doctorands’ Conference. Institute of
Economics, University of Veszprém 2003. pp. 66-72.
Komáromi Gy. (2003b): “Investment Decisions and Illusion of
Knowledge”. Competitio Vol. 2. No. 1. pp. 1-9.
Komáromi Gy. (2003c): “Psychological approach in Finance”, In: Borszéki
É. (ed.): Doctorands in Accounting and Finance. Department of Accounting
and Finance, Saint Stephan University 2003. pp. 75-81. (in Hungarian).
Komáromi Gy. (2004): “Was there a stock market bubble in Hungary?”
Competitio Vol. 3. No. 1. pp.169-178.
References 107
Komáromi Gy. (2004) Anatomy of Stock Market Bubbles. PhD

dissertation, University of Veszprém.
Kostolany, A. (1991): “Kostalanys Börsenpsychologie”. ECON Verlag
GmbN, Düsseldorf, Wien und New York. (in German).
Le Bon, G. (1982): The Crowd: A Study of the Popular Mind. Reprint
Fraser Pub.
Lei, V., Noussair, C. N. and Plott, C. R. (2001): “Non-speculative bubbles
in experimental asset markets: lack of common knowledge rationality vs.
actual rationality”. Econometrica Vol. 69. No. 4. pp. 831-859.
Lo, A. W., Mamaysky, H. and Wang, J. (2000): “Foundations of Technical
Analysis: Computational Algorithms, Statistical Interference, and Empirical
Implementation”. Journal of Finance Vol. 45. No. 4. pp. 1705-1765.
Major I. (2003): “The international crisis of infocommunication sector – an
institutional approach”. Információs Társadalom Vol. 3. No. 1. pp. 27-46. (in
Hungarian)
Malkiel, B. M. (1990): “Random Walk Down Wall Street” W.W. Norton
and Company, Inc. New York.
Medvegyev P. (2002): “Basic theorem of pricing financial asset in discrete
time models”. Közgazdasági Szemle, Vol. 49. Nos. 7-8. pp. 597-620. (in
Hungarian)
Morck, R., Yeung, B. and Yu, W., (2000): “The information content of stock
markets: why do emerging markets have synchronous stock price
movements?” Journal of Financial Economics Vol. 58. pp. 215-260.
Neal, L. (1992): “South Sea Bubble”. In: Newman, P. and al. (eds.): The New
Palgrave Dictionary of Money and Finance. III., Macmillan Reference,
London, 1992. pp. 492-495.
Odean, T. (1999): “Do Investors Trade Too Much?” American Economic
Review Vol. 89. pp. 1279-1298.
O’Quinn, R. P. (2003): “Economic Repercussions of the Stock Market
Bubble – a joint economic committee study”. United States Congress, USA.
July, 2003.
Pound, J. and Shiller, R. J. (1986): “Speculative Behavior of Institutional
Investors”. Journal of Portfolio Management, Vol. 13. No. 3. pp. 46-52.
Roll, R. (1984): “Orange Juice and Weather”. American Economic Review

Vol. 74. pp. 861-880.
Rotyis J. (2001): “Handbook for Investors on Stock Market” KJK-
KERSZÖV, Budapest. (in Hungarian)
Schwartz, R. A. (1992): “Market liquidity”. In: Newman, P. and al. (eds.):
The New Palgrave Dictionary of Money and Finance. II., Macmillan
Schwert, G. W. (1992): “Stock market crash of October 1987.” In:
Newman, P. and al. (eds.): The New Palgrave Dictionary of Money and
Finance. III., Macmillan Reference, London, 1992. pp. 577-582.
Shiller, R. J. (1981): “Do Stock Prices Move Too Much to Be Justified
Susequent Changes in Dividends?” American Economic Review Vol. 71. No.
3. pp. 421-436.
Shiller, R. J., (1991): “Market Volatility”. Third printing, The MIT Press
Cambridge, Massachusetts, London, England.
Shiller, R. J. (2000): “Irrational Exuberance”. Princeton University Press,
Princeton, New Jersey, USA.
Shleifer, A. (2000): “Inefficient markets – an introduction to behavioral
finance”. Oxford University Press. Oxford. UK.
Shleifer, A. and Summers, L. H. (1990): “The Noise Trader Approach to
Finance”. Journal of Economic Perspectives Vol. 4. No. 2. pp. 19-33.
Silber, W. L. (1992): “Secondary market”. In: Newman, P. and al. (eds.): The
New Palgrave Dictionary of Money and Finance. III., Macmillan Reference,
London, 1992. pp. 411-414.
Smith, V. L., Suchanek, G. L. and Williams, A. W. (1988): “Bubbles,
Crashes, and Endogeneous Expectations in Experimental Spot Asset
Markets”. Econometrica Vol. 56. No. 5. pp. 1119-1151.
Sornette, D. (2003): “Why Stock Markets Crash”. Princeton University
Press, USA.
Soros Gy (1994): “The Alchemy of Finance” Wiley, New York.
References 109
Srivastava, S. (1992): “Experimental asset market”. In: Newman, P. and al.

(eds.): The New Palgrave Dictionary of Money and Finance. I., Macmillan
Stanley, T. D. (1997): “Bubbles, Inertia, and Experience in Experimental
Asset Markets”. Journal of Socio-Economics, Vol. 26. No. 6. pp. 611-625.
Strathern, P. (2001): “Dr. Strangelove’s Game – A Brief History of Economic
Genius”. Penguin Books, London, England, 2002.
Szalavetz A. (2004): “Does The New Economy Bubble Burst?” Magyar
Tudomány, Vol. 49. No. 2. pp. 163-170. (in Hungarian)
Tarafás I. (2001): “The monetary policy from the great crises to unstable
money markets”. Aula Kiadó Kft., Budapest. (in Hungarian)
Thaler, R. H. (2000): “From Homo Economicus to Homo Sapiens”. Journal
of Economic Perspectives, Vol. 14. No. 1. pp. 133 141.
The Economist (1999a): “Hubble, bubble, asset-price trouble”. The World
Economy Survey. September 25th 1999. pp. 15-23.
The Economist (1999b): “Hot property”. Economics Focus. October 2nd
1999. p. 98.
The Economist (2003): “No pain, no gain”. Economic Focus. December
13, 2003. p.74.
Tirole, J. (1992): “Speculation”. In: Newman, P. and al. (eds.): The New
Palgrave Dictionary of Money and Finance. III., Macmillan Reference,
London, 1992. pp. 513-516.
Weller, P. A. (1992): “Rational bubbles”. In: Newman, P. and al. (eds.): The
New Palgrave Dictionary of Money and Finance. III., Macmillan Reference,
London, 1992. pp. 271-273.
White, E. N. (1990): “The Stock Market Boom and Crash of 1929
Revisited”. Journal of Economic Perspectives Vol. 4. No. 2. pp. 67-83.

Anatomy of Stock Market Bubbles

Uploaded by

Copyright:

Available Formats

Anatomy of Stock Market Bubbles

Uploaded by

Document Information

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

Anatomy of Stock Market Bubbles

Uploaded by

Copyright:

Available Formats

Anatomy of Stock Market Bubbles

ICFAI BOOKS is the initiative of the ICFAI University Press to publish

The articles are organized in a sequential and logical way that

Copyright © György Komáromi, 2006. All rights reserved.

First Edition: 2006

Published by The ICFAI University Press

ICFAI Editorial Team:

Dr. György Komáromi is Assistant Professor of Finance at CEU

I am also thankful to my colleagues at Pannon University and the

I am also very thankful to Krishna Chaitanya for giving me an important

I am grateful to my wife, my daughter and my family for their patience and

 List of Figures and Tables I

1. Different Aspects of Bubbles 1

2. Anatomy of Stock Market Bubbles 69

2. Anatomy of Stock Market Bubbles 69

3. The Hungarian Case, 1996-2003 88

“We know there are bubbles but we cannot even demonstrate

Economists often use the term “stock market bubble” to back up

I have first heard about the notion during my university studies,

I have considered claims like “shares are considerably overvalued

I was impressed by Burton Malkiel’s book, Random Walk Down

During my university years there were three books contributing

fluctuations”. John Kenneth Galbraith’s satirical essay A Short

The scientific question comes from the opening quote: how

My objective is to point out weaknesses in the main theories

The Subject and Structure of the Book

Studies classified by this distinction as mathematical use the

After this we will present the achievements of empirical

Although literary economics of a mostly historical-logical

Repeated, cyclical upsurge-downfall scenarios, however, bear

The second chapter of this paper will examine speculative

actors who base their investment decisions not on relevant

“Economics is a set of tools, from which the economist has to select

Charles Kindleberger (2000. p.220).

6 he phenomenon of stock market bubbles is explained in a

the investors’ expectations of further increase. This interpretation

After looking at the two-fold definition of bubbles, we will

1.1 Asset Price Bubbles in a Mathematical Approach

pt = Et (dt+1 + pt+1 )/(1+ r) (1.1)

where dt is dividend, pt is the price of the asset at t time, and

stochastic component satisfying the condition bt = Et(bt+1) /(1 + r),

Regarding stock investments, Keynes’ General Theory makes a

and if pt ≠ pt* then in the mathematical sense an asset price

This definition, however, needs some refinement. Formula 1.3

1.1.1 Rational versus Speculative Bubbles

other hand, it is a matter of speculative bubble if market price and

Asset price bubbles form in the presence of rational expectations

Rational bubble models determine an infinite number of

The weakness of the approach presented above is the implicit

Up to this point in our discussion, the existence of a bubble

rational but they base their decisions on differing (asymmetrical)

An expected bubble is to be construed as a scenario where actors

We are talking of a strong bubble when there can be no future

If, however, a market actor does not possess information in

Assuming rational expectations, therefore, share price will

Until now we have focused on the internal causes of bubble

the economic practicability of models integrating them, because

Studies assuming rational behavior teach us two important

1.1.2 The Problem of Determining Fundamental Value

Brunnermeier (2001. p.49) uses the following example to

List of Figures and Tables I

Table 1.1: Main Behavioral Patterns in Investors Decisions