1. Adam Smith Business School, University of Glasgow
In search of illegal
activity in the
Spanish stock market
MBA Programme – Final Dissertation
A dissertation submitted in part requirement for the
Master of Business Administration
JOSE FRANCISCO LAFRAGUETA PASCUAL
30-8-2016

2. 1
Table of contents
Acknowledgement ................................................................................................................................ 2
Abstract.................................................................................................................................................. 3
Declaration of originality .................................................................................................................... 4
Chapters:
1. Introduction ............................................................................................................................. 5
2. Literature review .................................................................................................................... 7
2.1 Mergers and Acquisitions (M&A) ................................................................................ 7
2.2 Run-up Hypotheses ...................................................................................................... 8
2.2.2 Insider Trading Hypothesis (ITH) ...................................................................... 9
2.2.3 Market Expectation Hypothesis .......................................................................... 9
2.3 Spain – Previous studies ............................................................................................. 12
2.4 Legislation on M&A ................................................................................................... 12
3. Hypothesis Development ...................................................................................................... 15
4. Data Methodology ................................................................................................................. 16
4.1 Data collection for target firms) .................................................................................. 16
4.2 Calculating Abnormal Returns (ARs) and CAARs ................................................... 17
4.3 Media Coverage study ................................................................................................ 21
4.4 Google Search Engine study ....................................................................................... 28
5. Findings .................................................................................................................................. 36
6. Discussion .............................................................................................................................. 40
6.1 Target firms sample .................................................................................................... 40
6.2 Media coverage study ................................................................................................. 40
6.3 Google search study .................................................................................................... 41
7. Conclusion ............................................................................................................................ 43
References ........................................................................................................................................... 45
Appendixes:
Appendix 1 Takeovers in the Spanish stock market during 2004-2016 .................................. 48
Appendix 2 Rumour on takeover activity................................................................................ 51
Appendix 3: Google search study - calculations...................................................................... 52
Appendix 4: Box-plot analysis................................................................................................. 53
Appendix 3: Regression analysis for AGoogle1...................................................................... 54
Appendix 4: Regression analysis for AGoogle2......................................................................55

3. 2
Acknowledgment
At this point, I would like to acknowledge some people that have especially contributed in the
achievement of this dissertation paper.
Firstly, I would like to thank my supervisor Evangelos Vagenas-Nanos for being always there
to help me solve any difficulties I met during the time spent on this thesis. Thanks for your
rapid responses and your willingness to help.
My sincere thankfulness goes to John Brady, a Ph.D. student at the University of Glasgow who
never let me down during the entire dissertation. He helped me to collect the data without which
this project would have never been possible. Always reachable despite being extremely busy
with his Ph.D. thesis.
I feel deeply thankful to my friend for their support all the way through, especially to Diego,
who helped me with the technical problems I had at the end of the project. I wish you all the
best with your upcoming exams.
I would like to thank my beloved girlfriend, and hopefully, future wife who has always been
there in the toughest moments to support me and care for me. At would like to extend my
thanks to my also beloved dog, Sam who has been with me every second and every moment I
have spent writing this dissertation. This paper is also yours.
Finally, I would like to specially mention my family, starting for my parent Jose Antonio and
Maria del Carmen who have supported me economically to achieve this MBA programme, and
also my brother who believes in me like no one, and has motivated me to study abroad. I am
proud of you, my family.
I would like to dedicate this dissertation to my grandmother Felicitas, who sadly passed away
two months ago. You taught me to never give up in life and fight for the things you want to
achieve. I will always follow your advice. Always with me, always in my heart.

4. 3
Abstract
We live in times of change, new business models and new technologies are disturbing the
apparent calm waters of our entire society. The world of Finance is not an exception, with
information being available for investors at levels like never before, and having an immediate
impact on stock markets, whoever possess information is bound to be in a superior position
than other investors in the stock markets. Power and money have always been associated with
corruption and illegal activities. This is disturbing and regulators try to balance by
strengthening rules and increasing the enforcement of legislation to avoid illegal activities.
In this new world, however, same old problems are faced every day by those regulators, how
to detect illegal activity amongst investors? How to measure the efficiency of the existing
legislation to protect those in disadvantaged positions? During almost 50 years, academia has
been trying to help in this field with no much success. But now, the disturbing technology is
offering a solution to those old problems; and researchers have been developing it for the last
six years with apparent robust results.
We are talking about Google, a seemingly inoffensive tool that hides priceless information for
those in search of illegal activity in the stock markets.

5. 4
Declaration of originality
Declaration of Originality Form
This form must be completed and signed and submitted with all assignments.
Please complete the information below (using BLOCK CAPITALS).
Name….JOSE FRANCISCO LAFRAGUETA PASCUAL ...................................................................
Student Number……2219296L...........................................................................................................
Course Name…MBA ..........................................................................................................................
Assignment Number/Name…...MGT5019 / MBA DISSERTATION....................................................
I confirm that this assignment is my own work and that I have:
Read and understood the guidance on plagiarism in the Postgraduate Handbook, including
the University of Glasgow Statement on Plagiarism 
Clearly referenced, in both the text and the bibliography or references, all sources used in the
work 
Fully referenced (including page numbers) and used inverted commas for all text quoted from
books, journals, the web etc. 
Provided the sources for all tables, figures, data etc. that are not my own work

Not made use of the work of any other student(s) past or present without acknowledgement

Not sought or used the services of any professional agencies to produce this work

In addition, I understand that any false claim in respect of this work will result in disciplinary
action in accordance with University regulations 
DECLARATION:
I am aware of and understand the University’s policy on plagiarism and I certify that this assignment
is my own work, except where indicated by referencing, and that I have followed the good academic
practices noted above
Signed .................................................................................................................................................

6. 5
Chapter1
1. Introduction
During the 20th
century, the rise of takeover activity across the business world drew the
attention of researchers to study target firm’s stock performances around the merger
announcement (Holland & Hodgkinson 1994; Gupta & Misra 1989). As a result, academic
studies confirm that the share prices of target firms show an increase during the weeks prior to
the public announcement, the so-called “targets’ share-price run-ups”.
Although consensus was reached amongst researcher on the share prices’ upward trend, a
debate arose regarding the reasons for such pattern. So far, two hypotheses have been
developed to explain it: Insider Trading Hypothesis (Keown & Pinkerton 1981); Market
Expectation Hypothesis (Jensen & Ruback 1983).
According to the former hypothesis, abnormal returns are due to trading on non-public
information. The corporate individual in possession of privilege undisclosed information about
the intended mergers, trade with corporate stocks before the announcement day to outperform
the market and gain abnormal returns. Note that the market is not aware of the information
which in turns is reflected in lower stock prices of the target firms’ share prices than expected
if that information was available for the market, thus the possibility of outperforming it. On the
other hand, trading expectation hypothesis argues that the market has instruments to predict
possible takeovers. For instance, expert market analysts, sharp investors, journalists, etc. that
are constantly looking for information or rumours to predict takeovers. When rumours around
such activities reach those market instruments, the market reacts accordingly, which in turn is
reflected in an increase in the target’s share prices before the announcement is made public.
Thus, it explains the share price run-ups.
Amongst the academic studies carried out during those years, especially in the US, Canadian,
and UK stock markets, there are some advocators of the market expectation hypothesis as the
unique explanation of the targets’ returns run ups (Pound & Zeckhauser 1990; Gupta & Misra
1989; Holland & Hodgkinson 1994). However, the majority of those studies claim that insider
trading is widespread and usually find that a combination of both hypotheses is the cause of
the increase in returns before the announcement (Jarrell & Poulsen 1989; Eyssell & Arshadi
1993; Gupta & Misra 1989). Since insider trading is considered illegal as it is against market
fairness, the findings of those studies were considered as robust evidence of the fragility of the
existing legislation to protect market efficiency. Thus, it outlined the necessity of new
legislation to protect investors, and better regulate stock markets.
The US was first in developing legislation to protect investors with a major reform of its
legislation by passing the Securities Act in 1933 and the Securities and Exchange Act in 1994
(Arshadi & Eyssell 1991). And the creation of a public body (SEC) to regulate the stock market
and prosecute insider trading activity. Yet, in the following years, several cases of insider
trading were found and prosecuted within the US market, which made the problem more
visible, and drew international attention (Thompson & Moines 2013). Therefore, with such
public demonstration of the existence of insider trading, other countries’ regulators were forced
to follow US example and enforce new legislation to protect stock markets worldwide.

7. 6
Nowadays, with a new bonanza in mergers and acquisition activity, the debate is back at the
front line of researcher studies (Cumming & Li 2011; Siganos & Papa 2015); and back to the
agenda of regulators (Bris 2005). In fact, a new European Directive to enforce market
efficiency across stock markets is being introduced by its country members, which is deemed
to produce significant changes. However, a common issue faced by both groups has always
been how to measure insider trading activity. On the basis that the cause of the target firms’
share price run-up is either of the two hypotheses, previous research studies find easier to
measure the market expectation hypothesis by focusing on the media coverage of intended
takeovers. Rumours present in financial newspapers, journals, magazines has always been
considered as a source of information for investors. Thus, researcher uses the presence of
rumours in them to measure the market awareness and assess their impact on share prices
(Mathur & Waheed 1995; Jabbour et al. 2000). If rumours cannot explain the run-ups, then it
is considered the presence of insider trading as a cause of such pattern.
This paper aims to bring more clarity to the existing share prices run-up debate by analysing
target firms’ abnormal returns in the Spanish stock market through the use of Google search
volume as a novel measure of investors’ attention.
In addition, this paper asks whether insider trading activity is present in the Spanish market by
assessing whether the market expectation hypothesis can explain the run-up pattern. We use
media coverage and Google search volume to demonstrate market awareness on intended
mergers, and compare results obtained by the two proxies. We follow this new approach that
has been developed by recent studies (Da et al. 2009; Siganos 2013) and applied in an event
study analysis of 60 Spanish takeovers recorded from 2004 and 2016.
The election of Spain as a subject to our study is due to the following reason: First, experts
consider that insider trading is widespread in this market (Navas 2015). Yet, little research has
been carried out to confirm this statement. Moreover, the entry of the new European Directive
in the Spanish legislation is considered to have an impact on insider trading activity. With
Spain recording low numbers of cases prosecuted in recent years (CNMV 2016), this paper
could measure the effectiveness of the previous law, and be used for comparison by future
research studies.
We contribute to the literature in several ways: Firstly, this is the first study to examine the
causes of the targets’ share prices run-ups in the Spanish market. To this author’s knowledge,
only Ocana et al. (1997) carry out a similar study on Spanish takeovers but fail to add any
explanation to the share prices upward trend discovered. Secondly, this paper’s contribution
extends to whether Google search volume is considered as a reliable proxy to measure
investors’ attention. Only a few studies have mastered this approach, showing positive results.
Finally, regulators can benefit from this study. Basically, if numbers can be allocated to
measure the level of insider trading in the market, legislation effectiveness could also be
measured. In fact, a new European Directive (2014) to prevent insider trading is to be amended
Spanish law in 2016.
The remaining of this paper is structured as follows: Section 2 describes the literature review,
with especial interest in previous studies regarding the two hypothesis. Section3 states the
research question. Section 4 presents the data methodology carried out during the event study
analysis which is divided into three sub-studies. Section 5 outlines the empirical results.
Section 6 describe the point of debate of the study. And, section 7 concludes.

8. 7
Chapter 2
2. Literature review
This chapter covers some key features involved in the merger and acquisition activity to better
understand this field. Starting from a more holistic perspective by analysing international
research studies, and international law, to then narrowing the scope to a more national level
since the Spanish stock market is the subject of our study.
2.1 Mergers and Acquisitions (M&A)
In today’s business world, an essential element of any corporate growth strategy is to enter into
emerging markets, and also to enhance market power in existing ones by means of acquisition
of other companies (Lebedow 2008). The main cause of this now common corporate behaviour
is Globalization (Bris 2005). However, it does not present a constant pattern in time, it reports
years where the number of M&A are high, followed by years of low scores, only to rise again,
and so on. This pattern is known amongst researchers as M&A waves (Harford 2005).
In the last 100 years, there have been six waves of rapid M&A activity. The first being at the
end of the 20th
century. Companies carried horizontal mergers in the same industries (steel and
oil) to increase market power through the creation of monopolies. The second wave came in
1960 with companies trying to diversify their offering within different sectors. In 1990, the
third wave was powered by the deregulation of some industries like transport or energy.
Finally, in the 21st century, globalization has been the driving force of the last three waves
involving industries such as banking and telecommunications (Lam 2015).
After the global crisis of 2008, which had a massive impact on the M&A activity, experts
believe that we are at the beginning of the seventh wave, with 2015 being the biggest year ever
for mergers and acquisition in the US market (Farrell 2015). This bonanza is supported by the
data collected from the Institute for Mergers, Acquisitions, and Alliances (IMAA) and shown
below in Figure 1. It can be seen the wave pattern across time.
As a result of the increase in M&A activity, investors’ attention is being boosted towards the
stocks of the companies involved in the takeover activity as a source of significant positive
abnormal returns. Hence, the demand for information on intended takeovers is now increasing
amongst individual investors, investing banks, pension funds, market stock market experts, and
even the media (financial journals). This high demand brings a market where information is
traded amongst those actors, with rumours on intended takeovers being common in the industry
on a daily basis. There are two types of information, public (rumours) and non-public
information (privilege insider corporate information) in this market. Trading with non-public
information is considered against the fairness of the market, as other investors cannot reach
that information. Thus, it is deemed illegal.
Therefore, regulators aware of this new M&A wave, are being advised to improve the existing
legislation in the stock markets to ensure their effectiveness (Bris 2005). Yet again, the old
debate about the Efficient Market Hypothesis (EMH) is being challenged. Is it possible to

9. 8
outperform the market, and gain abnormal returns when investing in stocks of takeover target
companies?
Figure 1: Mergers and Acquisitions worldwide
The figure shows the number of takeovers successfully close worldwide during the last 20 years. Date collected
from IMAA (https://imaa-institute.org/statistics-mergers-acquisitions/)
2.2 Run-up Hypotheses
Within the process of mergers and acquisitions, it is worldwide known that bidder firms pay
large premiums to take control of target companies (Schwert 1996). Thus, the share prices of
target companies experiment high increases on the announcement day as the market reacts and
adjusts itself to the public information of the agreed premium to be paid, and other benefits that
are believed to increase the value of the target company due to synergies. This adjustment on
share prices aligns with the Efficient Market Hypothesis (EMH) which postulates that prices
always reflect the available information (Oberlechner & Hocking 2004). The possibility of
gaining substantial abnormal returns from these premiums has stimulated investors’ interest in
obtaining information on potential takeover operations (Zivney et al. 1996) which in turn has
an impact on share prices during the weeks/months before the announcement, the so-called
“stock price run-ups”.
In the field of Finance, many academic studies confirm that target’s share prices do increase
previous the formal announcement of the merger (Jabbour et al. 2000; King 2009; Mathur &
Waheed 1995; DeAngelo et al. 1984). With target’s run-ups studies reporting cumulative
abnormal returns (CARs) ranging from 5 to 32.25% over the 30-50 days up to the
announcement date, with 50% of the increase happening before the very announcement date
(Jabbour et al. 2000; Meulbroek 1992). Therefore, the market reaction to possible mergers
starts to occur before the announcement day, which in turns generates an upward trend in the
share prices, increasing as the announcement date approaches. In order to explain the reason
for this pattern, a debate has been active amongst researchers since the early 80s, developing

10. 9
two hypotheses to explain such pattern: Market Expectation Hypothesis; and Insider Trading
Hypothesis.
2.2.1 Insider Trading Hypothesis
Insider Trading Hypothesis (Keown & Pinkerton 1981), people involved in the negotiations of
the possible mergers, so called “insiders”, use this private information to benefit themselves by
buying shares of these firms to gain from the expected premium. It is the first study that uses
daily returns to analyse the abnormal returns in a sample of 194 firms in the American Stock
Market. Basically, the study is carried out on the assumption that if the market is efficient and
all the public information is reflected in the market share price, only those with inside
information can outperform the market. Results show what it was considered common
knowledge at that time, that information about intended mergers are poorly held secrets and
trading on this non-public information abounds due to leakage of information from insiders.
Later in time, other studies back-up these initial findings, Meulbroek (1992) finds that around
half of the increases in target’s share prices occurs on insider trading days which suggests that
insiders use this private information during those trading days to buy/sell shares of target
companies. The Same finding is reported in later studies (Schwert 1996). Moreover, King
(2009) finds evidence on the existence of a pattern between non-public or illegal trading with
the increase in abnormal returns, and also trading volume activity previous public
announcement. Jabbour et al. (2000) also present evidence of insider trading in their sample of
128 takeovers in the Canadian market during 1985-1995, showing CARs of 12.28% in the 61-
days run-up period before the announcement.
These studies highlight the necessity of legislation to prevent trading on non-public information
in order to keep fairness and clarity amongst investor in the Financial Markets. The US Market
was the pioneer in introducing this type of legislation through the Security Exchange Act in
1934 and later amended in 1984 by the Insider Trading Sanctions Act (ITSA), which applies
severe penalties to insiders who leak private information on mergers and acquisitions, and
regulated by the Security Exchange Commission (SEC) (Arshadi & Eyssell 1991). In recent
years, the increase of mergers and acquisition due to “globalisation” has brought stricter
legislation changes and an increase in enforcement, not just in the US market but all financial
markets. However, since illegal insider trading is difficult to demonstrate no consensus
amongst researchers has been agreed on the effectiveness of the legislation framework to
prevent illegal activity in this field (Arshadi & Eyssell 1991; Aktas et al. 2008; Bris 2005;
Linciano 2003; Eyssell & Arshadi 1993), claiming that this type of illegal trading is still present
at different levels around the world.
2.2.2 Market Expectation Hypothesis
Market Expectation Hypothesis (Jensen & Ruback 1983), is based on the believe that investors
predict the firms that will become targets before the companies make the announcement public,
based on information released on news, magazines, corporate reports, dividend changes,
regulatory changes, etc. Researchers claim that information is available to anticipate future
mergers through skilled investors, analyst advisors, the so-called “shark watcher” who are part
of the system to make the market efficient (Oberlechner & Hocking 2004; Bhabra 2008).

11. 10
Early studies based on the US and Canadian markets use media coverage (newspapers) as a
proxy to measure the market’s awareness of intended mergers or acquisitions before the public
announcement (Pound & Zeckhauser 1990; Zivney et al. 1996; Jabbour et al. 2000; Kiymaz
2001). Results are contradictory as the information published in the newspapers and treated as
“rumours” for possible takeovers cannot explain the totality of the CARs experienced in those
studies, suggesting that a combination of the two hypothesis is the most rationale answer to
explain the target’s share price run-ups.
For instance, Pound & Zeckhauser (1990) uses information published in the Wall Street
Journal’s HOTS column in a small sample of 42 firms during the period of 1983-85, findings
that few rumours were caused by leaks from insiders, instead they are caused by close
observation of unusual activity in firms’ stocks by professionals as evidence of possible
mergers. However, they find abnormal returns of 7.78% previous the publication of the
rumours which might evidence illegal activity. In addition to this, Jarrell & Poulsen (1989) find
that although public news explains a significant part of the run-ups prior announcement for
firms in the news, there is also evidence of substantial pre-news increase on the target’s share
prices. On the other hand, Zivney et al. (1996), in a study that also analyses the HOTS column
and the AOTM column, argue that the market reacts efficiently and no evidence of insider
trading in a sample of 871 rumours related to takeovers published during 1985-88.
Similar results are gathered from studies carried out in international markets, a combination of
both hypotheses is the most sensible approach to explaining the target’s share run-ups (Siganos
& Papa 2015; Kiymaz 2001). An early study in the UK market that also uses rumours within
news as a proxy for market anticipation hypothesis finds no evidence of insider trading in a
sample of 86 target firms from 1988 to 1989 (Holland & Hodgkinson 1994). Whereas, in a
most recent and complete study of the UK market, Siganos & Papa (2015) analyse a wider
sample of 783 target firms from 1998 to 2014 by using Financial Times (FT) coverage of
rumours around 60 days previous the announcement of those mergers. They find that the media
coverage can only explain 27% of the run-ups in firms with previous rumours. Therefore, the
findings can be considered as strong evidence of the existence of insider trading which the
authors argue it can be linked with the presence of a softer UK legislation compared to the US
market for instance.
Few studies have been carried out in smaller international markets to bring more clarity to the
hypothesis debate. At the Istanbul Stock Exchange market, Kiymaz (2001) follows the “media
coverage” approach on a sample of 355 rumours mentioned in HOTS column of local financial
newspaper “Ekonomik Trend”. Results benefit the presence of illegal trading as pre-
publication run-ups are significant. Although, the author suggests that findings can be biased
as the column HOTS usually mention stocks that usually have recently been performing well.
However, contrasting the studies that attribute pre-bid share price run-up of target firms to
insider-trading behaviour, a studied carried out in another emerging market – Australian
market- concludes that no significant pre-bid run-ups are found after considering a broad range
of public information sources in a sample of 450 takeover offers between the years 2000 and
2009 (Aspris et al. 2014).
All the previous studies highlight a common problem when using the “rumours in media”
approach to explaining the reason of the targets’ share price run-ups, the difficulty of gathering
all the public information available to analyse the efficiency of the Market Expectation

12. 11
Hypothesis. Most of them are based on rumours mentioned in a specific publication such as
Financial Times or similar publication in each country. Therefore, the findings can be
considered biased as they do not represent all the information available at that particular time,
investors have other sources of information to predict intended mergers such as forums,
conferences, direct conversations, extreme returns, etc. (Siganos 2013).
Another problem related to all these sources of information available for investors to predict
mergers is that researchers use them as a measure of investors’ attention on a particular stock
to explain different patterns or behaviours, but these sources are only indirect proxies of
attention (Barber & Odean 2008). The common assumption is that if a stock was mentioned in
the media or had extreme returns, then investors should have paid attention to it. However, the
simple presence of rumours in media does not guarantee attention if investors does not actually
read the article, and extreme returns are driven by more factors unrelated to investor attention
(Da et al. 2009). However, as technology developed, a novel and direct proxy to measure
investors’ attention / market awareness have recently been developed to help understand the
target firms’ run-ups pattern – Google Search Volume (Da et al. 2009; Siganos 2013).
In their study, De et al. (2009) argue that the aggregate search frequency in Google is a reliable
measure of investors’ attention for several reasons: First, a search engine is commonly used by
internet users to collect information, and Google is considered the favourite amongst them.
Second, if somebody is searching for a stock in Google is directly paying attention to it.
Therefore, aggregate search frequency is an unambiguous measure of attention. Finally,
economists have already used it to describe public interest in a variety of economic activities
such as home sales, tourism, etc. The findings of this study in US firms reveal strong evidence
that Google search frequency captures the attention of individual investors earlier than existing
indirect proxies, such as media coverage. Authors claim that an increase in the frequency
predicts higher stock prices in the short-run (next 2 weeks), especially for small stocks. The
results are also supported by the finding of a later study carried out in the German market. Bank
et al. (2010) suggest that an increase in Google search frequency is associated with an increase
in stock trading volume and future stock returns, and also associated with reduced stock
liquidity. Thus, it demonstrates the usefulness of search data in financial applications.
To some extent, in line with these two previous studies, Siganos (2013) also supports the use
of Google search volume as a proxy for investor attention. Although, his study focuses on the
use of this novel proxy to analyse the target firms’ run-ups before their public announcement
to demonstrate the presence of illegal insider trading in the UK market. The assumption is that
when investors receiving rumours or hints on potential mergers, they will search for further
information on the target firm stock on the Internet via Google search before initiating any
transaction. Therefore, firms, that feature in rumours captured in any of the mentioned indirect
sources of information available for investors, are expected to show an increase in Google
search in a specific moment of their historic activity. Thus, this increase in Google search
activity could be considered as the specific time when the market is aware of the intended
merger, and therefore we could calculate if the targets’ run-up trend follows the Market
Expectation Hypothesis, or to the contrary the Insider Trading Hypothesis.
In a sample of 430 UK’s target companies in mergers occurred during 2004-2010, Siganos
(2013) reports that, although, Google indicators predict a larger percentage of the price run-ups
in target firms than the media coverage proxy (Financial Times), they only explain the 36% of

13. 12
it. Thus, the paper shows evidence of insider trading in the UK market, the author suggests that
the cause for this phenomena is the presence of a softer M&A legislation compared to other
countries. This paper follows Siganos’ approach to explore Spanish takeover activity.
2.3 Spain-Previous studies
To this author’s knowledge, there is only one research paper analysing takeover targets’ returns
in the Spanish stock market following the methodology of the previously mentioned
international studies (Ocana et al. 1997). The paper examines a sample of 71 target companies
involved in takeovers recorded from 1990 to 1994. They find abnormal positive returns for the
target firms with a significant upturn within the two months before the bid announcement.
However, in line with studies in larger markets (US or UK), the highest increase in cumulative
abnormal returns is located within the 30 days before the announcement day.
In fact, the total share price increase reported is much higher than other papers, 41% with one-
third of the total amount being earned prior to the official takeover. This result suggests that
shareholders of target companies gain significant premiums in the Spanish market, higher than
other larger markets at that time (Ocana et al. 1997).
However, unlike previous international studies, the authors do not extend their study to analyse
the causes of this pattern. Neither the Market Expectation Hypothesis or the Insider Trading
Hypothesis are mentioned in the paper as a possible explanation for such upward trend in target
abnormal returns. Therefore, it can be said that there is a substantial gap in the research area of
this phenomena in the Spanish market. Hence, this paper aims to fulfil some of the existing
gaps in research of Spanish takeover activity.
Another study carried out in the Spanish stock market that we consider worth mentioning in
this section is Del Brio et al. (2002). The authors examine insider trading activity and its
probability. They find evidence of insiders earning returns that exceed risk-adjusted
benchmarks by using their private information on corporate prospects. The result of this study
question the effectiveness of Spanish law against insider trading and recommend a regulatory
change to prevent this illegal activity. Moreover, they also find that outsider cannot earn
abnormal profits when information made public.
2.4 Legislation on M&A
The Insider Trading Hypothesis is based on the presence of insider trading activity of share,
bonds, derivatives, and other instruments in the financial markets. Insider trading occurs when
individuals with potential access to non-public corporate information buy and sell stocks of
that company to outperform the market. In this cases, individuals make use of their privilege
positions to gain access to non-public information that other investors cannot be aware of. Such
trading is considered illegal (Thompson & Moines 2013). Takeover activity is one of the fields
where this type of activity can generate major benefits to insiders due to the premiums paid for
the bidder company to target company’s shareholders. Thus, it should be legally protected.
Regulation and reinforcement against insider trading are important for investors mainly for
three reasons: Firstly, investors are prone to be more confident in financial information released

14. 13
by companies operating in countries with strong legislation in place to avoid insider trading.
Secondly, investments are deemed to be less risky in those countries as information is
considered more reliable. And third, investments are likely to require lower rates of return as
risk and required a return by investors are positively correlated.
Historically, the United Stated has been considered to be the leading force in insider trading
law. This is not a surprise considering that the US has the two largest stock exchanges in terms
of capitalization, NASDAQ, and New York Stock Exchange (NYSE). Thus, the US ranks first
in the world by market capitalization with a total value around $26,000 billion (The Word
Bank, 2016). Therefore, legislation in insider trading is really important.
In the US, the first regulation implemented was the Securities Act in 1933, and the Securities
and Exchange Act in 1934. Both Acts were approved to increase transparency to investors by
increasing the requirements on securities’ public information. By means the second Act, the
Security and Exchange Commission (SEC) was created to regulate the trading of securities.
The SEC in combination with the Department of Justice has the power to create and enforce
the rules to regulate the securities market. The SEC defines inside trading as, “… any person
has violated any provision of this title or the rules or regulations thereunder by purchasing or
selling a security or security-based swap agreement … while in possession of material, non-
public information…” (SEA, 1934). It implies that anyone with access to non-public
information and acting on it can be convicted of insider trading.
In fact, since the implementation of the law, many cases have been brought to court by the SEC
against directors, employees, government officers, etc., who gained millions of dollars by
trading on confidential information with penalties of a maximum of 20 years in prison, and
fines up to $5 million (Thompson & Moines 2013). This rather large record of cases highlights
the existence of this illegal activity no just on the US market, but all over the world. In the last
20 years, there has been a global commitment to enforce insider trading law amongst
international countries, however, it is difficult to judge results.
The Spanish stock market (BME) with a capitalization value of $900 billion is much smaller
than the US market or in European terms, the UK market (around $3,000 billion). However, it
is considered a big market, which ranks within the top 15 largest worldwide (The World Bank,
2016). Legislation regarding securities was almost non-existent before Spain joined the
European Economic Community in 1986. As a result, the Spanish Securities Market Act (1986)
was implemented, completely reforming the financial market.In addition to this, the Comision
Nacional del Mercado de Valores (CNMV) was created as the Spanish version of the US’s
SEC to regulate the market. The Act defines insider as anyone who possess insider information,
with insider information being defined as, “information of a precise nature, relating to one or
more issuers of securities which have not been public” (Thompson & Moines 2013). Listed
firms have the legal requirement to report the CNMV of any firm-related event that could have
a significant impact on market prices. In the case of not fulfilment of this duty, and found guilty
of insider trading activity only economic fines are applied, no prison penalties for the individual
involved. Thus, it defers from the US’s legislation.
The previous lack of regulation had an impact on investors’ confidence. In fact, the first
takeover, or in Spanish defined as Oferta Publica de Adquisicion (OPA), was registered in 1983
(Ocana et al. 1997). The implementation of the Act has had a positive impact on the market,
with the number of OPAs being increased scientifically. However, nowhere near the number

15. 14
of M&A registered for bigger markets such as the US, or the UK. For instance, during the
period 1991-2002, the CNMV’s records show that 142 OPAs were registered in Spain; and 103
since 2004 until 2016 (CNVM, 2016). On contrast, UK registered 170 takeovers only in
2015(Office for National Statistics, 2016).
During the last decade, there has been a debate about the effectiveness of the European
regulation in insider trading. The amount of cases brought to court in the European countries
is considered “ridiculous” when comparing with the US. In fact, Spain recorded only 7 cases
prosecuted for insider trading activities from 2000 and 2006, with experts claiming that trading
with non-public information is common practice across investors in Spain (Navas, 2015).A
statement also backed-up by the findings of research studies (Del Brio et al. 2002; Ocana et al.
1997). Both papers find evidence of insider trading activity in the Spanish stock market.
Other research papers find evidence of this illegal behaviour across other European countries,
such Italy and UK (Linciano 2003; Siganos & Papa 2015). Both papers claiming the need of
an increase in toughness and strictness of the law to reduce insider trading activity. Hence, the
European Commission is preparing a new Directive 2014/57/UE, known as the Market Abuse
Regulation (MAR), to enforce regulation across its members, including Spain.
Amongst others, the main change that the new legislation brings is the inclusion of prison
penalties. A maximum of 4 years-penalty for those with insider trading charges; and 2 years
for those involved in leaking no-public information to others. Another change is the increase
in the economic fines to be applied to those involved. These changes aim to level European
law with the US law in this field, which is considered more effective (Linciano 2003).

16. 15
Chapter 3
3 Hypothesis development
This paper aims to bring more clarity to whether insider trading is a common practice in the
Spanish stock market by carrying out an empirical analysis of share prices of target companies
involved in takeovers, or OPAs during the last decade. In order to deliver our objective, the
following hypothesis is developed:
First, we have to find whether takeover activity recorded during 2004 and 2016 reports share
price run-ups amongst the target companies involved. This is the base of our study, since the
targets’ return run-up has to be explained by either of the two hypothesis. Basically, if market
expectation hypothesis can explain the totality of the share prices upward trend, then insider
trading is considered no to have an influence in this field. Thus, previous legislation is to be
deemed effective and the low number of cases of insider trading taken to court would be
explained. On the contrary, if the market expectation hypothesis cannot completely explain the
upward trend, then insider trading is deemed to be the cause, which in turn would undermine
the effectiveness of the existing legislation. To understand the process, it is worth mentioning
that share price run-ups are formed by daily abnormal returns. This paper follows the market
adjusted returns method to obtain abnormal returns:
ERiM = Rit – RMt, where Rit is the return of firm “i” in day “t”, and RMt is the market
return in day”t”.
Ones the run-up trend is measured; we follow the market expectation hypothesis to explain
such trend. We follow two different proxies in this stage. First, in line with the early studies in
the field, we use media coverage (takeover rumours in financial journals) as a proxy of
investors’ attention. In order to calculate the impact of these rumours on the share prices, we
have to collect the dates when those rumours were published and run an event analysis. If the
market expectation is to explain the trend, results should show no share-price run-up before
rumours. On the contrary, the difference between the two run-ups is considered the percentage
of the former trend is explained by the information available in the market (public information).
The rest is considered to be due to illegal insider trading (non-public information).
The second proxy and the core part of this study is the use of the novel Google search volume
as investors’ attention. It follows the same approach than the media coverage study. In line
with previous studies, we expect that this proxy can explain a larger part of the former run up
than the media study, if not all. Therefore, this paper is to add more empirical data to confirm
the usage of Google as an effective alternative to discover insider trading activity in the stock
market during merger and acquisitions activity.
Measuring insider trading is deemed to be one of the major limitation regulators are facing to
assess legislation effectiveness. Therefore, if this paper’s findings support the use of Google
search volume as a more accurate and effective method to measure insider trading activity,
Regulators could benefit from its use by comparing legislation effectiveness in different
periods. Thus, to better understand whether enforcement is necessary or not.

17. 16
Chapter 4
4. Methodology
This chapter explains the methodology followed in the event study to explore in more detail
the target firms’ share price run-ups. The paper focuses on the mergers and acquisitions (OPAs)
occurred in the Spanish Market during the period 2004-2016. The aim is to bring more clarity
to the debate of what hypothesis explains better the run-up pattern, MEH or ITH.
The study is divided into three major stages. The first stage focuses on data collection where
daily return are needed to calculate both, the abnormal returns (ARs), and the cumulative
abnormal returns, which in turn are the components to show the run-up trend. Secondly, in line
with most of the studies in this field, this study uses the “rumours in the media” proxy to justify
the MEH as a reason for the uptrend. Third, the study uses the novel “Google search” proxy to
demonstrate the validity of this new approach to measuring the market awareness of intended
OPAs in the Spanish Market.
4.1 Data collection for target firms
I use Thomson OneBanker to download information of Spanish target firms involved in
completed OPAs between 1/1/2004 and 1/07/2016. The election of the period of study is
limited by the use of the Google search engine, which provides data collected after 2004, and
the starting date of this study. Other requirements of the sample are:
i. Firms that own at least 50% of the target company after transaction
ii. Public firms (registered in the Spanish Stock Market)
iii. Firms should be assigned with unique ticker symbols
iv. Exclude reverse takeovers
v. Exclude self-tender takeovers
vi. Exclude recapitalization transaction
Moreover, to be included in the sample, firms should have available DataStream codes to link
Thomson OneBanker with DataStream. DataStream is used to download target’s daily share
returns. The ticker symbol requirement will be explained in more detail later in this paper, in
the section that deals with the Google search engine proxy.
The process yields a final sample of 60 target firms. It is believed to be a small sample if
comparing to most recent studies which use similar approaches. For instance, Siganos (2013)
works in a study of 430 UK firms from 2004 and 2010. And Siganos & Papa (2015) analyse
one of the largest studies so far, with a sample of 783 UK firms. Another large sample is carried
out in the Canadian Market with 399 firms (King 2009). However, this study’s sample aligns
with other previous studies: 87 US firms (Gupta & Misra 1989); 86 UK companies (Holland
& Hodgkinson 1994); 42 US firms (Pound & Zeckhauser 1990). It is worth noticing that this
study is carried in a smaller market as it is the Spanish Stock Market compared to the US,
Canadian and UK markets, and also the period of time chosen includes the years of the global
economic crisis which are believed to have had an impact on the number of mergers occurred

18. 17
during those years. Therefore, it can be considered a valid sample that can bring significant
findings and clarity to the hypotheses debate.
Appendix 1 shows the table with the information collected from the Thomson OneBanker on
those 60 target companies. In addition to this Table1 (below) shows a summary of the number
of OPAs completed in each year of the sample period.
Table 1: Summary of OPAs during the sample period
Year Number of OPAs Year Number of OPAs
2004 1 2011 5
2005 4 2012 10
2006 5 2013 5
2007 6 2014 7
2008 3 2015 5
2009 6 2016* 1
2010 2 TOTAL 60
* Until 1/7/2016
The table shows that the Spanish Market is not a really active market in terms of mergers and acquisition with an average of 5
completed OPAs a year. The lowest numbers are registered in 2008 and 2010, with 3 and 2 OPAs respectively, which verifies
that the global economic crisis had an impact on the number of mergers completed during those years. 2012 onwards, it shows
signs of recovery with the higher number of OPAs being completed.
Once the sample of target firms is fixed, we use DataStream to collect the daily returns of those
companies to proceed with the analyse of the share price run-ups. As mentioned before, a
DataStream code for each company is necessary to download the required information. Once
the daily share price returns of the whole sample are collected, further calculations are needed
to represent the targets’ run-ups.
4.2 Calculating Abnormal Returns (ARs) and CAARs
I use the daily stock return data collected from DataStream to calculate daily abnormal returns
(ARs) for each firm around the period of the announcement by means of the market adjusted
returns model. So, given that the market return for each day (Spanish market index) is collected
from the DataStream, the formula for our study is the following expression:
Where: is the abnormal return of stock “i” on day “t” is the difference between
the return of stock “i” on day “t” and the market return.

19. 18
Once the ARs are calculated for the sample period around the announcement, the next step is
to calculate the Average Abnormal Return (AAR) for each day of the sample period (day 0,
day -1, day -2, … day -30) over the total number of firms (N). Calculations follow the following
mathematical expression:
Finally, the Cumulative Average Abnormal Return (CAARs), defined as the sum of previous
daily averages is also calculated for each trading day of the study. In this study, different blocks
of CAARs are used to analyse the run-up trend which will be explained in following chapters.
For this particular point, the CAARs are calculated for each day of the sample period following
this expression:
Where: is the sum of the average abnormal returns (AAR) between day
“t1” and “t2”
The common belief is that if there are no unusual share price activity prior to the announcement
day, one would expect that AARs and CAARs would fluctuate randomly around zero. On the
contrary, if there is a leakage of information, and trading on securities occurs prior to the
announcement day, positive daily average returns should appear as “t” approaches day 0,
showing an increased build-up in CAARs.
In line with previous studies (Gupta & Misra 1989; Meulbroek 1992; Jabbour et al. 2000;
Jarrell & Poulsen 1989; Spyrou et al. 2011), this paper analyses the daily returns around the
commonly used period of 30 days prior to the announcement, allocating day 0 to the
announcement day. Some other studies use a wider study period. For instance, Siganos & Papa
(2015) study a 60-day period prior to announcement day, however, they also find that the
uptrend starts to occur around day -35. Similar results are found in another study set over the
60-day period before the announcement day in the US market (Keown & Pinkerton 1981).
They find that the CARs become positive 25 trading days prior to the announcement.
In addition to this, the only study carried out in the Spanish market before this paper finds
results in line with previous studies in the US and Canadian markets, with rising CAARs
around a day -25 (Ocana et al. 1997). This results can be considered significant since the study
period of this research paper is around 260 days prior to announcement day. Therefore, this
paper considers the period (-30,0), as the correct horizon to analyse the possible target’s share
price run-up in the Spanish Stock Market during 2004 and 2016. Results for the calculations
are shown in Table 2 and represented graphically in Figure 1 and Figure 2 (below).

20. 19
Table 2: Daily Average Abnormal Returns (AARs) and Cumulative Average Abnormal
Returns (CARs) surrounding the announcement day (0).
This table presents the AARs and CAARs surrounding the announcement day (0) for targets in Spanish takeovers. AARs and
CAARs are calculated following the formulas explained previously.
Figure 2: Average Abnormal Returns (AARs) for target companies in Spanish takeovers
This figure shows the AARs for 60 target companies traded on the Spanish Stock Market during 2004 and 2016. ARs were
calculated using the market model, where the stock market return is given by the IGBM index.
Event day AARs CAARs Event day AARs CAARs
-30 0.00% 0.00% -14 0.69% 0.60%
-29 0.28% 0.28% -13 -0.38% 0.23%
-28 0.19% 0.47% -12 -0.02% 0.21%
-27 0.17% 0.65% -11 0.07% 0.28%
-26 0.21% 0.86% -10 0.30% 0.58%
-25 0.07% 0.93% -9 -0.32% 0.26%
-24 -1.82% -0.89% -8 0.04% 0.29%
-23 0.44% -0.45% -7 0.19% 0.48%
-22 -0.10% -0.56% -6 0.47% 0.96%
-21 -0.23% -0.79% -5 0.24% 1.19%
-20 0.57% -0.22% -4 0.25% 1.45%
-19 0.11% -0.11% -3 0.04% 1.48%
-18 -0.12% -0.23% -2 0.79% 2.27%
-17 -0.16% -0.40% -1 0.91% 3.18%
-16 -0.17% -0.56% 0 2.93% 6.11%
-15 0.48% -0.08%
-2.00%
-1.00%
0.00%
1.00%
2.00%
3.00%
4.00%
-30-29-28-27-26-25-24-23-22-21-20-19-18-17-16-15-14-13-12-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
ARs
Day
Average Abnormal Returns
mtiiitit RRAR  ˆˆ 

21. 20
Figure 3: Cumulative Average Abnormal Returns (CARs) for target companies in
Spanish takeovers.
This figure shows the CAARs for 60 target companies traded in the Spanish Stock Market during 2004 and 2016. CAARs
were calculated based on .
The empirical results show a clear upward trend for the target firms’ share prices prior to the
announcement day (0) in the Spanish Stock Market during the sample period (2004-2016),
which is in line with previous studies on the field of targets’ share price run-ups in bigger Stock
Markets, such as the US or UK. In our sample, AARs fluctuate normally till one week before
the announcement day when an increase in returns is noticed. From day -8 to day -7, returns
increase from 0.04% to 0.19%, only to show a higher increase in the day -7 with returns of
0.49%. As we approach day 0 values keep increasing with day-2 and -1 scoring 0.79, and
0.91%, respectively. The upward trend is complete on day 0 when the share prices reach their
higher returns with a 2.93%. It might suggest a higher volume of trading on the sample target
firms’ shares which in turn suggests that investor might have been aware of the intended
mergers days before the announcement.
The results for the CAARs shown by figure 2 confirm the share price run-up trend perform for
the AARs for our sample. However, the upward trend commences on the day -15 compared to
the one shown by the AARs on the day -7. Before day -30 to day -15, CAARs perform with
negative returns. From day -15 onwards till the announcement day, the trend is always positive.
It shows a constant increase from 0.23% on a day -16 to the final 6.11% on day 0, which means
that the share price total aggregate returns for the study period (-30,0) are 6.11%. It is also
worth noticing that with the final CAR (-30,0) recording 6.11% includes the increase in AAR
registered on the announcement day, 2.93%. If we consider that the information is already in
the market that very day, it suggests that around a 52% of the total share price run-up occurs
-2.00%
-1.00%
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
-30-29-28-27-26-25-24-23-22-21-20-19-18-17-16-15-14-13-12-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
CARs
Day
Cumulative Average Abnormal Returns

22. 21
on previous days to the official announcement day. The result is similar than other papers’
findings: 36% (Holland & Hodgkinson 1994); 52% (Keown & Pinkerton 1981); 44% (Jarrell
& Poulsen 1989); 38% (King 2009); between 40-50% (Meulbroek 1992)
Therefore, and following the approach of previous papers (Holland & Hodgkinson 1994), it
reduces the unexplained run-up study period to one between day -30 to day -1, which shows a
Cumulative Abnormal Return of 3.18%. Although lower than other studies, this result is still
in line with most previous studies which find similar CAARs for similar samples in other Stock
Markets: 4.0 % in the Spanish Market (Ocana et al. 1997); 3.6% in UK (Spyrou et al. 2011);
9.30% in UK (Siganos 2013); 7.5% in US (Zivney et al. 1996); 7% in US (Pound & Zeckhauser
1990); and 5.7% in Canada (King 2009). Hence, it can be said that the findings are robust and
that the study sample of target firms for takeovers represents evidence of the existence of share
price run-ups in the Spanish Stock Market. Thus, the study of this sample could be relevant to
bring clarity to the debate about the cause of this upward trend in share prices between the
Market Expectation Hypothesis and the Insider Trading Hypothesis.
This paper tries to explore the former hypothesis by exploring whether the target share price
run-up is driven by public information available in the market. Previous papers use the
“rumours in the media” proxy to explain the trend. This paper uses the same proxy to assess
whether the market was aware of the intended mergers before the announcement day, and
therefore to explain the run-up. Next section covers this point.
4.3 Media coverage study
It commonly recognized that one of the sources of information for investors is financial
newspapers. They tend to cover the latest news in the financial markets, such as rumours on
possible mergers and acquisitions (M&A). Hence, it has been used as a market attention proxy
for past researchers to demonstrate the Market Expectation Hypothesis in the target firms’ share
price run-ups (Jarrell & Poulsen 1989; Siganos & Papa 2015).
This paper uses the same approach to demonstrate whether the market was aware of the
intended takeovers before the bid announcement in our sample of 60 Spanish target companies,
which in turn might explain all or part of the 3.18% increase in share price returns recorded
during the period (-30, -1), with the announcement day being day 0.
As mentioned in the literature review, some researchers use specific journals’ columns where
rumours could be found in their studies. For instance, Pound & Zeckhauser (1990) use the Wall
Street Journal’s “Heard on The Streets” column; the same column is used by Zivney et al.
(1996), modern studies use a wider approach analysing all the articles in the Financial Times
journal, as technology helps with a more rapid means of collecting information (Siganos &
Papa 2015).
For my study, this paper makes use of a Spanish newspaper called “Cinco Dias”, which is
deemed to be the most used newspaper amongst investors in Spain to gather information, and
also the worldwide known Financial Times. The selection of these two newspapers is for
robustness porpoises. The small size of the Spanish Stok Market could be a constraint for an
international journal as it is the Financial Times, and it is thought that small takeovers do not

23. 22
be covered by it. Hence, the use of a national newspaper “Cinco Dias” to ensure whether or not
rumours of intended Spanish takeovers were available in the press.
I use the Nexis to access daily coverage of both newspapers, “Cinco Dias” and “Financial
Times”. I search for articles for target firm names and set the time span around three months
before each takeover announcement. It is worth noticing that the search in undertaken using
full text instead of only headlines which yield a higher number of relevant articles matching
the wanted firm name. Articles found by Nexis may not be relevant to rumours on takeover
activity involving the company, therefore each article is carefully read to ensure the article is
relevant in our study. The process is carried out twice for each of the 60 target firms of our
study, firstly with articles in the “Cinco Dias” journal, and secondly with the articles published
in the “FT” journal to ensure relevance in the findings. Appendix 2 shows an example of a
published article that can be considered as a rumour for a possible takeover.
Table 3 shows the result of the media coverage for the sample study. The first column reflects
the name of the target company involved in the OPA activity, with the announcement day of
the official bid shown in the second column. The third column reads the value in millions of
dollars of the transaction. The dates of the first article found about the intended OPAs is
recorded in the fourth column and highlighted in bold font. We find 38 takeovers with rumours
before the bid is made official, meaning that a 63% of the takeovers from our sample have had
some type of media coverage, either national (Cinco Dias) or international (FT). To the
contrary, 37% of the mergers were not covered by the media. This is to some extent in line
with previous studies, for instance, Siganos & Pala find 25% of the UK mergers of their study
with no FT coverage. For these mergers with no rumours found, the announcement date is kept
as the first rumour published. The last column in our table shows the difference in days from
the first rumour to the formal bid offer. The study finds rumours being published in average 23
days before the announcement day, which could be considered as noticeable since the study
period for the target run-ups has been fixed to the month before the official bid announcement
(-30, -1).
However, since the time span used for the rumour search is three months before the
announcement day, it could be interesting to allocate the results to each month. Monthly groups
are described as follows: 3rd
month (-90, -61); 2nd
month (-61, -31); and 1st
month before
announcement (-30, -1). I find that 6 takeovers are already covered by the media at least during
the 3rd month. 13 takeovers of the study sample start to be mentioned by the media (journals)
during the 2nd
month before the formal bid day. Yet, 19 intended takeovers are mentioned in
the media during the 1st
month, which coincides with the study period of the run-up.
To sum up, although the overall average of the rumours suggests that information on takeovers
is normally covered by media around 21 days before the announcement day (1st
month), only
19 takeovers from 60 (31%) are found during that 1st
month, and therefore only those 19
takeover rumours might be considered to have a real impact on describing the increase in AARs
during the study period. The media coverage yield by these 19 rumours recorded during this
1st
month is around -12 day, which means that usually rumours tend to appear two weeks before
announcement day.
The other 69% of the rumours is found within the previous two months of the study period. It
evidences that some takeovers require of long negotiation periods before final acceptance,
which gives media more time to predict them. They are usually takeovers with high valuations.

24. 23
This is confirmed in our study, from the 6 takeover rumours collected in the 3rd
month, 4 of
them happen to be high valuation ones (+ $1,000 mil). It is also worth noticing that media
coverage of this long negotiation does not ensure the acceptance of these deals, therefore it
might not impact the share prices in the same manner than those rumours found on days closer
to the announcement, where there may be more certainty about the success of the negotiations.
Table 3: Rumours on the Spanish takeovers collected from the media coverage search
Target Name
Date
Announced
Value
($mil)
Date of 1st
published
rumors Source
rumors
before
day 0
(days)
Testa Inmuebles en Renta SA 08/06/2015 971.95 17/04/2015 Cinco Dias -52
Gen de Alquiler de Maquinaria 31/03/2015 11.97 31/03/2015 - 0
Bodegas Bilbainas SA 22/01/2015 3.14 22/01/2015 - 0
Funespana SA 29/12/2014 24.82 29/12/2014 - 0
Sotogrande SA 17/10/2014 287.14 17/10/2014 - 0
Grupo Tavex SA 26/09/2014 17.49 26/09/2014 - 0
Cementos Portland Valderrivas 21/05/2014 151.99 27/03/2014 Cinco Dias -55
Ahorro Familiar SA 17/12/2013 34.85 17/12/2013 - 0
Campofrio Food Group SA 14/11/2013 966.20 24/09/2013 Cinco Dias -51
Dogi International Fabrics SA 02/10/2013 5.15 31/08/2013 Cinco Dias -32
Banco de Valencia SA 04/04/2013 33.51 09/01/2013 Cinco Dias -85
Corp Dermoestetica SA 20/12/2012 3.68 20/12/2012 - 0
Metrovacesa SA 19/12/2012 1,063.42 16/10/2012 Cinco Dias -64
Cia d'Aigues de Sabadell SA 19/12/2012 31.88 03/12/2012 Cinco Dias -16
Banesto 17/12/2012 346.62 23/11/2012 Cinco Dias -24
Secuoya Grupo de Comunicacion 27/09/2012 20.30 27/09/2012 - 0
Banca Civica SA 26/03/2012 1,305.19 04/02/2012 Cinco Dias -51
Banco Pastor SA 07/10/2011 1,465.37 07/10/2011 - 0
CAM 14/07/2011 3,961.52 17/06/2011 FT -27
Telvent GIT SA 31/05/2011 1,583.00 31/05/2011 - 0
Befesa Medio Ambiente SA 17/03/2011 164.59 26/02/2011 Cinco Dias -19
Iberdrola Renovables SA 08/03/2011 2,132.14 08/03/2011 - 0
CEPSA 16/02/2011 4,964.33 16/02/2011 - 0
Banco Guipuzcoano SA 25/06/2010 4,964.33 12/06/2010 FT -13
Iberia Lineas Aereas de Espana 12/11/2009 2,674.04 09/09/2009 FT -64
Agbar 22/10/2009 419.22 30/09/2009 Cinco Dias -22
Banco de Andalucia SA 19/05/2009 2,904.27 19/05/2009 - 0
Vueling Airlines SA 09/01/2009 1,308.66 07/01/2009 Cinco Dias -2
Grupo Ferrovial SA 19/12/2008 3,999.61 19/12/2008 - 0
Banco de Credito Balear SA 25/09/2008 144.98 20/09/2008 Cinco Dias -5
Banco de Galicia SA 25/09/2008 49.79 20/09/2008 Cinco Dias -5
Banco de Castilla SA 25/09/2008 44.37 20/09/2008 Cinco Dias -5
Banco de Vasconia SA 25/09/2008 18.06 20/09/2008 Cinco Dias -5
Union Fenosa SA 14/08/2008 10,283.58 25/07/2008 FT -20

25. 24
Reyal Urbis SA 06/08/2008 156.73 30/06/2008 Cinco Dias -37
CELO 24/07/2008 12.38 24/07/2008 - 0
Logista 25/01/2008 1,398.45 12/11/2007 Cinco Dias -74
Sogecable SA 20/12/2007 3,050.59 21/11/2007 Cinco Dias -29
Plarrega Invest 2000 SA 23/10/2007 7.12 23/10/2007 - 0
Uralita SA 03/09/2007 12.57 03/09/2007 - 0
Parquesol SA 23/07/2007 1.09 05/05/2007 FT -79
Endesa SA 02/04/2007 26,437.77 03/01/2007 Cinco Dias -89
Altadis SA 14/03/2007 17,872.72 10/02/2007 FT -32
Riofisa SA 19/01/2007 2,576.19 16/01/2007 Cinco Dias -3
FADESA Inmobiliaria SA 28/09/2006 4,444.32 28/09/2006 - 0
Europistas CESA 04/08/2006 907.75 30/06/2006 Cinco Dias -35
Inmobiliaria Urbis SA 28/07/2006 4,085.59 24/07/2006 Cinco Dias -4
Inmobiliaria Colonial SA 06/06/2006 2,605.04 27/05/2006 FT -10
Telefonica Publicidad 28/04/2006 3,654.46 15/04/2006 FT -13
Telefonica Moviles SA 16/03/2006 4,214.07 16/03/2006 0
Tele Pizza SA 20/02/2006 718.10 20/02/2006 - 0
Hullas del Coto Cortes SA 23/12/2005 35.25 23/12/2005 - 0
Cementos Lemona SA 02/12/2005 307.09 07/11/2005 Cinco Dias -25
Grupo Inmocaral SA 05/07/2005 217.01 19/05/2005 Cinco Dias -47
Cortefiel SA 20/06/2005 1,518.88 12/05/2005 FT -39
Cie Automotive SA 06/06/2005 136.56 06/06/2005 - 0
Terra Networks SA 10/02/2005 5,821.59 18/12/2004 Cinco Dias -54
Aldeasa SA 27/01/2005 1,020.42 15/12/2004 FT -43
Recoletos Grupo Comunicacion 14/12/2004 1,251.52 02/11/2004 FT -42
Centros Comerciales Carrefour 31/05/2004 173.96 29/05/2004 FT -2
This table shows the media coverage on the 60 Spanish takeovers occurred between 2004 and 2016. Two newspaper were
used as a source of rumours, the Spanish journal “Cinco Dias” and the international journal “Financial Times” (FT). Data was
collected using Nexis database.
Ones the dates of the first rumours published in the newspapers have been collected, this paper
also studies the impact of these rumours on the target firms’ abnormal returns and whether or
not they can explain the run-ups. In order to proceed, we run the same study used before to
discover the run-up trend, but using the dates of the first rumours rather than the announcement
day, if rumours found. Otherwise, we keep the day of the official bid. To avoid confusion, we
name it “Media coverage” study. The hypothesis behind it is that if rumours are found before
the day of the announcement (day 0), then they might be part of the CAR (-30, 0) run-up trend,
which was previously highlighted to be 6.11 %. Since the media coverage is meant to be part
of the run-up trend, the closer to day - 30 the lower CARs the study is supposed to yield. By
comparing results from both studies, it can be calculated how much of the target firms’ run-
ups is due to public information (Market Expectation Hypothesis), or non-public information
(Insider Trading Hypothesis).
Table 4 (below) summarises the findings after having run the study. When analysing the data,
we find that the on day of the announcement (day 0), the Average Abnormal Return (AAR) is
2.97 in the original study, which makes 48 % of the entire run-up during the period month of

26. 25
the day of the official bid. This is considered as evidence that the market reacts efficiently to
the information when made public. In the case of the media coverage study, the day of the first
published rumour (day 0), the AAR is 2.06. out of the 5.46 for CAR (-30, 0). This is considered
as also evidence that the market reacts efficiently to rumours, and in turn, it gives a sense of
robustness to our findings from the rumours search study.
When looking at the AARs, it can be seen that in the last 15 days before day 0, the media
coverage study records more negative values (7), than the initial study which only scores 3
negative values during this period.
It is also noticeable that when looking at the daily Cumulative Average Abnormal Returns
(CAARs) in the media coverage study, they fluctuate randomly during the first two weeks of
the study period, only to show a constantly increasing trend after day -8 whereas, as mentioned
previously, in the initial study they turn positive and show the rising trend after day -15.
Comparing the CAARs (-30, 0) from both periods, 6.11 is recorded for the initial study and
5.46 from the media coverage study, which means that the media coverage can only justify
10% of the initial study’s share price run-up. Following the idea that the information is already
in the market the day before of publications (Holland & Hodgkinson 1994), we decided to
ignore day 0 from our study period, only to find a surprising result. CAAR (-30, -1) from the
media coverage study is 3.40, which is superior to the CAAR for the same period in the initial
study (3.18). Thus, the media coverage study can justify 0% of the initial run-up.
In order to better understand this rather unexpected result, we decided to separate this study
period in three blocks of 10 days each, (-30, -21); (-20, -11); (-10, -1). The media coverage
study scores 0.80; -0.13; and 2.73, respectively. This data compared to the initial study’s data
yields contrasting results. If we focus on the latest 10-days block, where the run-up is more
pronounced, the media coverage can only explain 0.05% of the initial run-up. In addition to
this, we also create a new block (-20, -1), which captures the total run-up period of the initial
study (starting around day -16). The 20-day block scores a 2.60 which comparing with 3.97
from the initial study in the same period, it suggests that around 34% of the target’s run-up
period is justified by the release of rumours by the media in days before the announcement day.

27. 26
Table 4: Media coverage study
This table shows a comparison between the initial target’s share price run-up study and the later media coverage study. By
comparing the AARs and CAARs from both studies during the same time span, this paper tries to explain the share price run-
up pattern.
Event day AARs CAARs AARs CAARs
-30 0.00% 0.00% -0.55% -0.55%
-29 0.28% 0.28% 0.07% -0.48%
-28 0.19% 0.47% 0.08% -0.40%
-27 0.17% 0.65% 0.79% 0.39%
-26 0.21% 0.86% 0.98% 1.37%
-25 0.07% 0.93% 0.27% 1.64%
-24 -1.82% -0.89% -0.75% 0.88%
-23 0.44% -0.45% 0.20% 1.09%
-22 -0.10% -0.56% -0.24% 0.85%
-21 -0.23% -0.79% -0.05% 0.80%
-20 0.57% -0.22% 0.41% 1.21%
-19 0.11% -0.11% -0.36% 0.86%
-18 -0.12% -0.23% 0.00% 0.86%
-17 -0.16% -0.40% 0.05% 0.91%
-16 -0.17% -0.56% -0.18% 0.73%
-15 0.48% -0.08% -0.28% 0.45%
-14 0.69% 0.60% -0.06% 0.39%
-13 -0.38% 0.23% -0.04% 0.35%
-12 -0.02% 0.21% 0.18% 0.53%
-11 0.07% 0.28% 0.14% 0.67%
-10 0.30% 0.58% -0.07% 0.60%
-9 -0.32% 0.26% -0.43% 0.17%
-8 0.04% 0.29% -0.17% 0.00%
-7 0.19% 0.48% 0.44% 0.45%
-6 0.47% 0.96% 0.41% 0.85%
-5 0.24% 1.19% 0.31% 1.16%
-4 0.25% 1.45% -0.01% 1.15%
-3 0.04% 1.48% 1.32% 2.47%
-2 0.79% 2.27% 0.60% 3.07%
-1 0.91% 3.18% 0.33% 3.40%
0 2.97% 6.11% 2.06% 5.46%
CAARs (-x , -y)
(-30, -21) -0.79% 0.80%
(-20, -11) 1.07% -0.13%
(-10, -1) 2.90% 2.73%
(-30, -1) 3.18% 3.40%
(-20, 1) 3.97% 2.60%
(-30, 0) 6.11% 5.46%
Announcement day Media coverage
Announcement day Media coverage

28. 27
As results from the full sample study are inconsistent, this paper decides to follow the approach
taken by Siganos & Papa (2015) and present result separately. The full sample is divided among
the companies which present media coverage of their merger (38), and the companies with no
media coverage (22).
Table 5 show results from this subdivision following the same structure of the former media
coverage study. If we focus on the run-up period excluding the day of the announcement,
CAAR (-30, -1), the firms with media coverage experiment the highest target price run-up,
with a 4.13 vs 2.46 for the firms with no media coverage. It suggests that rumours have an
impact on the share price upward trend. This is in line with previous studies (Siganos & Papa
2015).
Table 5: Subdivision study of media coverage
This table shows AARs and CAARs for our sample firms with media coverage and without media coverage. Then results are
compared with the full sample study.
Within the period (-20, -1), firms with rumours still presenting higher returns, recording 3.97
vs 2.91 for firms with no rumours. However, the next subgroup (-10, -1) presents a change over
previous results, now is firms with no rumours the ones scoring higher returns, 3.85 vs 2.90.
This change is later confirmed on the announcement day, with the firms with no rumours
scoring almost double abnormal returns than firms with rumours, 3.25 vs 1.36, which differs
from the findings from Siganos and Papa’ study, where firms with coverage have the strongest
share price reaction at the time of the announcement. Moreover, it is also worth noticing that a
Event day AARs CAARs AARs CAARs AARs CAARs
-10 0.30% 0.58% -0.25% 0.03% 0.07% 1.13%
-9 -0.32% 0.26% -0.57% -0.54% -0.08% 1.05%
-8 0.04% 0.29% -0.33% -0.87% 0.14% 1.20%
-7 0.19% 0.48% 0.65% -0.23% 0.28% 1.48%
-6 0.47% 0.96% 0.52% 0.30% -0.03% 1.45%
-5 0.24% 1.19% 0.64% 0.93% -0.10% 1.34%
-4 0.25% 1.45% 0.23% 1.16% -0.09% 1.25%
-3 0.04% 1.48% 1.99% 3.15% 0.25% 1.50%
-2 0.79% 2.27% 0.91% 4.06% 0.13% 1.63%
-1 0.91% 3.18% 0.07% 4.13% 0.83% 2.46%
0 2.97% 6.11% 1.36% 5.49% 3.25% 5.71%
(-30, -21) -0.79% 1.22% 0.77%
(-20, -11) 1.07% -0.94% 0.29%
(-10, -1) 2.90% 3.85% 1.39%
(-30, -1) 3.18% 4.13% 2.46%
(-20, 1) 3.97% 2.91% 1.69%
(-30, 0) 6.11% 5.49% 5.71%
Full sample Media coverage No media coverage

29. 28
similar share price run-up for both subgroups for the total study period (-30,0) is yielded, 5.49
for firms appearing in the media, and 5.71 for firms with no media coverage.
Despite finding evidence to support the impact of rumours on the appearance of targets share
price run-ups, media coverage does not explain the patterns entirely since firms with no media
coverage still present an upward trend before the announcement day. However, it supports the
findings from the previous full sample study which finds that media coverage can explain about
34% of the study sample share price run-up.
Hence, the use of media coverage as a market awareness proxy presents some limitations to
bring more clarity to the debate between the two run-up hypotheses. It is worth remembering
that this study applies two journals to find rumours on intended mergers in the Spanish Stock
Market, however, there are many more newspapers covering financial activities, and therefore
it can be used as a source of rumours. Similar constraints are found by previous researchers in
this field (Pound & Zeckhauser 1990; Oberlechner & Hocking 2004; Bris 2005). More recent
studies (Da et al. 2009; Siganos 2013) have developed a novelty new proxy to measure the
market awareness which is deemed to be more acquired to explain the run-up pattern – the use
of Google search engine as a direct measurement of investor attention. This paper applies this
new direct proxy to explain the increase of returns in the share prices of the 60 OPAs found
during 2004-2016 in the Spanish market. Next section covers the process followed and the
results collected by this new study.
4.4 Google Search Engine study
Google Trend website provides data on the frequency that any particular term has been
searched for during a set period of time. Data is available from back to January 2004, which in
turn delimitates our sample period, OPAs from 2004 to 2016. Google Trends measures the
search frequency via its Search Volume Index (SVI). SVI is a relative value to the total search
in a requested period of time, which ranges between 100 and 0, with 100 being the
day/week/month with more searchers, and 0 the one with less. The unit of the measure depends
on of the length of the requested study period. For instance, if the study period is three years,
the SVI data is given in months, with 100 being assigned to the month with more searchers.
Whereas, if the study period is shorter, 1 year, the SVI data is given weekly. Google Trend
gives the option to delimitate your study period, we use the three-month period before the
merger announcement, which yields SVI daily data. In addition, it coincides with the time span
used for the media coverage study, which allows comparison between both studies; and more
importantly, covers the time period where the targets’ run-up is present.
A major concern is the identification of a stock in Google. There is two option available: The
first option is to use the name of the target company, and the second option is to use its ticker
name. Using the company name as a search term could be problematic for two basic reasons:
Investor can be searching for not investing reasons, for instance, ones may be searching for
Coca-Cola because of the launch of a new marketing campaign rather than collecting financial
information about the company. Another issue related to the company name option is that when
searching for a company, different people may search for different names. For instance, in the
previous example, some will use Coca-Cola, some Coke, some CocaCola plc, etc. However,
despite the mentioned drawbacks, the company names option is used by a previous study on

30. 29
investors’ attention and trading volume carried out in Germany (Bank et al. 2011). They claim
that the use of company names may capture the attention that the firm is receiving for a much
broader audience since it seems unlikely that the average Google user would searcher by mean
of ticker names.
On the other hand, searching for stocks by using ticker names reduces ambiguity. If an investor
is searching for a ticker name (“APPL” for the company Apple Computer Inc.) in Google, it is
likely he/she is looking for financial information about Apple’s shares. This option is chosen
by Da et al. (2011), in what is considered the first study using Google search as investor’s
attention proxy to measure targets’ share price run-ups. Siganos & Papa agree on the benefit
of ticker names upon company names in their most recent study on UK companies. This papers
follows the same ideas than those authors and considers the ticker option as the most reliable
approach. Hence, when collecting the initial sample of the 60 Spanish target companies from
the Thomson One database, we set the requirement of the existence of a unique ticker name
allocated to each company. Therefore, our sample already contains the data necessary to
proceed with the Google Search Engine study.
For each firm, we perform the following process, the first step is to run the search engine with
the ticker name during the 3-month time span before the announcement date, which was already
collected to perform the first run-ups study explained in the first section of this chapter. Google
trends respond showing a graphical representation of the daily Search Volume Index (SVI) for
the required time period. Appendix 4 contains an example of one of the graphs collected during
the study representing the volume search activity for the selected ticker name. Besides the
graph, Google Trends also permits to download the daily SVI data in a CVL format file, we
download each file for each firm’s ticker name (60 firms) to an Excel file to analyse the
information.
One the data is downloaded to an Excel file; the next step is to estimate Abnormal Google
search returns. The hypothesis behind the study is that Abnormal returns in Google volume
activity represent changes in investor attention, which in turn might be driven by new
information about the intended mergers being collected by the investor. If the Google volume
search activity equals investors’ attention on a particular stock, it is believed that the Google
activity is to increase as we get closer to the announcement day since information is more likely
to be available on days close to the day of the official bid, and attract more attention of savvy
investors looking for intended mergers to gain generous returns on stock trading.
For robustness purposes and in line with the research study carried out by Siganos (2013), we
use two different measures to find daily Abnormal Google volume changes:
AGoogle1i = ln (1+SVIit) – ln (1+SVIit-1)
AGoogle2i = ln (1+SVIit) – ln [median (1+SVIit-31, 1+SVIit-32,…,1+SVIit-40)]

31. 30
Where STVit is the Google activity of firm “i” on day “t”, which for estimation purposes is
adjusted to a range between 1 and 2. Google1i and Google2i are estimated daily from day -30
to the announcement day (day 0) to capture the target price run-ups. Google1i records the daily
changes in search volume by difference between SVI values on day “t” and day “t-1”.
Google2i shows the abnormal daily change over the normal Google activity for each target
firm. It is estimated by the difference of the SVI value on day “t” and the median number of
searchers between the day -31 and day -40 before the merger, which is considered as
representative of the normal Google activity.
To confirm the robustness of the data, we calculate the Cumulative Average SVI for the entire
sample (60 target firms) for each day during the period between day -30 and day 0
(announcement day) for both measurements, AGoogle1i, and AGoogle2i. Figure 3 and Figure
4 show the results obtained from the robustness test. It can be seen that both AGoogle1i and
AGoogle2i tend to increase towards the announcement day, with both recording an increase of
0.23, and 1.30%, respectively during the 10 days before the announcement day. Also both
scoring positive abnormal volume of searchers on the day of the official announcement, 0.20
and 1.42%, respectively. It suggests a positive relationship between Google volume and shares
returns. This is in line with findings from previous studies(Da et al. 2009; Siganos 2013).
Figure 4: Robustness test of AGoogle1
Figures 4 shows the distribution of AGoogle1 searchers across the study period, 30 days before the announcement day

32. 31
Figure 5: Robustness test of AGoogle2
Figures 5 shows the distribution of AGoogle2 searchers across the study period, 30 days before the announcement day.
Ones the SVI data is being confirmed as representative of investors’ attention for our sample,
we proceed with the study to demonstrate that Google Search Index can be used to explain the
targets’ share price run-up pattern by calculating what day investor are aware of the intended
mergers. The process is similar to the one followed by the “media coverage” study carried
earlier in this paper, where the date of first rumour reflects the market awareness of the intended
mergers. In this case, the date recording the first significant Abnormal SVI is taken as the day
that reflects the market awareness, therefore comparing the study of the Abnormal Returns
between the Google study and the initial announcement day study will demonstrate if the
Market Expectation Hypothesis can explain the target run-up.
To identify significant upward changes across the studies, we follow the outlier literature since
both measures, AGoogle1i and AGoogle2i are continuous variables. To proceed, we first
explore the distribution of both measures, recording the results in Table 6 (below). The table
reads that both measures are positively skewed, 0.92 and 0.52, respectively. With AGoogle1i
recording a Kurtosis pick of 1.16, vs a negative Kurtosis value recorded by AGoogle2i of -
0.18. Both measures do not follow a normal distribution at 1%, therefore we follow the boxplot
method to identify outliers (Tukey, 1977).
This method is also carried by Siganos (2013), as presents similar data statistics. By using this
methodology, we discover if the value score for AGoogle1/AGoogle2 for the firm “i” on any
particular day “t” from the study period can be really considered as Abnormal from the rest of
the value obtained, and therefore be considered as the day investors are aware of the
information

33. 32
Table 6: Descriptive statistics of Google search measurements
The table shows descriptive statistics for both measures, AGoogle1i and AGoogle2i which are generated from Google Trends
data.
Ones decided the methodology to calculate the outliers, calculations are carried across the study
sample of 60 target companies. The Boxplot methodology generates the following formula:
Outliers > Q3i + 1.5 * (Q3i – Q1i)
Where Q3i and Q1i are upper and lower quartiles for firm “i” over the period between -30 days
and the day of the announcement. The first outstanding abnormal upward change amongst
ASVI values for each firm is considered the first signal of an intended merger activity.
Appendix 3 shows an example of these calculation being carried out for a specific target
company. A first and second column of the table read the event day ranging from -90 to day 0
representing the 3-month time span of the initial ticker search. shows the SVI data downloaded
in the first step. The third column present the SVI data collected from Google Trends. 4th
and
5th
columns show the calculation for Google1, and the last two columns show the calculation
for Google2. An addition graph is generated as a visual aid to analyse the data. An additional
table is created to show calculations for the identification of outlier amongst the data from
Google1 and Google2. If an outlier is identified during the study, the date matching the value
with the day of the study is collected, otherwise, the announcement day is maintained as the
day investors are aware of the information. We find that in most cases, the identified outlier
matches the pick value of the period study. Finally, findings are double-checked by the use of
SPSS software, which provides the option of calculating bot-plot analysis (Appendix 4).
After running the calculations for the 60 target companies of our sample, a summary is created
to resume the findings, Table 7 (below). The first three columns in the table read the known
information about the target companies and the announcement days. In the 4th
and 5th
columns,
the dates corresponding to the outliers found during the data analyses are highlighted in bold,
otherwise, the announcement day is kept. We find 49 outliers for AGoogle1 and 29 for
AGoogle2. The last two columns show the difference between the day of the announcement
and the outlier. AGoogle1 average a value of 10.7, and AGoogle2 average a value of 6.7. This
is to say that the former option predicts the information about intended mergers around 11 days
AGoogle1i AGoogle2i
Average 1.90% 4.59%
Median 1.62% 3.83%
Min -10.89% -7.25%
Max 22.98% 19.97%
Standard Deviation 0.07 19.97%
Skewness 0.92 0.52
Kurtosis 1.16 -0.18

34. 33
before the announcement day, and for the latter option is around day 7 before the takeover in
our sample study.
However, to truly identify how much of the targets’ share price run-up can be explained by this
Google Search Engine study, the dates of the identified outliers (Abnormal SVIs) are to be used
to calculate the AARs and CAARs for the targets’ share prices of the study sample. Then, by
means of comparison between results we can estimate an answer. Next paragraph covers this
process.
Table 7: Resume of findings from Google search study
Target Name
Ticker
Symbol
Date
Announced Google 1 Google2
days
before bid
date
(Google1)
days
before bid
day
(Google2)
Testa Inmuebles en Renta SA TST 08/06/2015 29/05/2015 31/05/2015 10 8
Gen de Alquiler de Maquinaria GALQ 31/03/2015 03/03/2015 31/03/2015 28 0
Bodegas Bilbainas SA BBI 22/01/2015 13/01/2015 22/01/2015 9 0
Funespana SA FUN 29/12/2014 29/12/2014 29/12/2014 0 0
Sotogrande SA SOTG 17/10/2014 18/09/2014 18/09/2014 29 29
Grupo Tavex SA TVX 26/09/2014 26/09/2014 14/09/2014 0 12
Cementos Portland Valderrivas CPL 21/05/2014 19/05/2014 21/05/2014 2 0
Ahorro Familiar SA AHOF 17/12/2013 17/12/2013 17/12/2013 0 0
Campofrio Food Group SA CPF 14/11/2013 11/11/2013 23/10/2013 3 22
Dogi International Fabrics SA DGI 02/10/2013 17/09/2013 02/10/2013 15 0
Banco de Valencia SA BVA 04/04/2013 04/04/2013 04/04/2013 0 0
Corp Dermoestetica SA DERM 20/12/2012 10/12/2012 20/12/2012 10 0
Metrovacesa SA MVC 19/12/2012 17/12/2012 19/12/2012 2 0
Cia d'Aigues de Sabadell SA AIG/B 19/12/2012 10/12/2012 19/12/2012 9 0
Banesto BTO 17/12/2012 21/11/2012 20/11/2012 26 27
Secuoya Grupo de
Comunicacion SEC 27/09/2012 08/09/2012 08/09/2012 19 19
Banca Civica SA BCIV 26/03/2012 06/03/2012 17/03/2012 20 9
Banco Pastor SA PAS 07/10/2011 11/09/2011 18/09/2011 26 19
CAM CAM 14/07/2011 09/07/2011 26/06/2011 5 18
Telvent GIT SA TLVT 31/05/2011 18/05/2011 31/05/2011 13 0
Befesa Medio Ambiente SA BMA 17/03/2011 14/03/2011 15/03/2011 3 2
Iberdrola Renovables SA IBR 08/03/2011 07/02/2011 08/03/2011 29 0
CEPSA CEP 16/02/2011 07/02/2011 16/02/2011 9 0
Banco Guipuzcoano SA GUIP 25/06/2010 18/06/2010 11/06/2010 7 14
Iberia Lineas Aereas de Espana IBLA 12/11/2009 24/10/2009 24/10/2009 19 19
Agbar AGS 22/10/2009 12/10/2009 22/10/2009 10 0
Banco de Andalucia SA AND 19/05/2009 19/05/2009 19/05/2009 0 0
Vueling Airlines SA VLG 09/01/2009 18/12/2008 29/12/2008 22 11
Grupo Ferrovial SA FER 19/12/2008 30/11/2008 30/11/2008 19 19
Banco de Credito Balear SA CBL 25/09/2008 25/09/2008 25/09/2008 0 0
Banco de Galicia SA GAL 25/09/2008 25/09/2008 25/09/2008 0 0
Banco de Castilla SA CAS 25/09/2008 08/09/2008 25/09/2008 17 0
Banco de Vasconia SA VAS 25/09/2008 29/08/2008 25/09/2008 27 0

35. 34
Union Fenosa SA UNF 14/08/2008 27/07/2008 29/07/2008 18 16
Reyal Urbis SA REY 06/08/2008 17/07/2008 27/07/2008 20 10
CELO - 24/07/2008 24/07/2008 24/07/2008 0 0
Logista LOG 25/01/2008 14/01/2008 25/01/2008 11 0
Sogecable SA SGC 20/12/2007 02/12/2007 20/12/2007 18 0
Plarrega Invest 2000 SA PLI 23/10/2007 23/10/2007 23/10/2007 0 0
Uralita SA URA 03/09/2007 03/09/2007 03/09/2007 0 0
Parquesol SA PSL 23/07/2007 16/07/2007 23/07/2007 7 0
Endesa SA ELE 02/04/2007 02/04/2007 02/04/2007 0 0
Altadis SA ALT 14/03/2007 25/02/2007 25/02/2007 17 17
Riofisa SA RFS 19/01/2007 19/01/2007 19/01/2007 0 0
FADESA Inmobiliaria SA FAD 28/09/2006 24/09/2006 26/09/2006 4 2
Europistas CESA EURM 04/08/2006 18/07/2006 18/07/2006 17 17
Inmobiliaria Urbis SA IURE 28/07/2006 25/07/2006 25/07/2006 3 3
Inmobiliaria Colonial SA ICLE 06/06/2006 06/06/2006 06/06/2006 0 0
Telefonica Publicidad TPI 28/04/2006 17/04/2006 28/04/2006 11 0
Telefonica Moviles SA TEM 16/03/2006 01/03/2006 13/03/2006 15 3
Tele Pizza SA TPZ 20/02/2006 11/02/2006 19/02/2006 9 1
Hullas del Coto Cortes SA HCC 23/12/2005 04/12/2005 26/11/2005 19 27
Cementos Lemona SA CPDC 02/12/2005 20/11/2005 22/11/2005 12 10
Grupo Inmocaral SA MOC 05/07/2005 05/07/2005 05/07/2005 0 0
Cortefiel SA CTF 20/06/2005 20/06/2005 20/06/2005 0 0
Cie Automotive SA CIEA 06/06/2005 02/06/2005 03/06/2005 4 3
Terra Networks SA TRLY 10/02/2005 28/01/2005 28/01/2005 13 13
Aldeasa SA ALD 27/01/2005 02/01/2005 02/01/2005 25 25
Recoletos Grupo Comunicacion REC 14/12/2004 19/11/2004 19/11/2004 25 25
Centros Comerciales Carrefour CRF 31/05/2004 24/05/2004 31/05/2004 7 0
The table shows the dates corresponding the Abnormal SVI from the Google Search study. They are calculated by means of
the outlier methodology, which follows the equation: Outliers > Q3i + 1.5 * (Q3i – Q1i)
In the final step of this study, the dates identified as outliers or abnormal upward change
(investors ‘attention), are used in an event study analysis to calculate prior Abnormal Returns
and Cumulative Abnormal Returns. The hypothesis behind this study is that if the market is
efficient, and the dates found as outliers suggest the day when the information was publicly
available, then the calculations should yield no Abnormal Returns amongst the study sample.
Or at least, present a smoother targets’ share price run-up. Thus, the difference between both
studies’ CAARs (-30, -1) is to be the amount of the trend demonstrated by public awareness,
or in other words the Market Expectation Hypothesis. On the contrary, the value for CAAR (-
30, -1) from the new study is to be considered the abnormal returns generated from non-public
information trading, in other words, Insider Trading Hypothesis.
Table 8 presents the results generated when running the outlier dates collected from the Google
Search Engine study. For analysing purposes, it also reads the results obtained when running
the announcement days (former study). Results are explained in the next paragraph.

36. 35
Table 8: Target share returns generated in the Google Search Engine study
The table reports Average Abnormal Returns (AARs) and Cumulative Average Abnormal Returns (CARs) generated from the
dates identified as the moment where the market was aware of intended mergers from the Google search study.
Event day AARs CAARs AARs CAARs AARs CAARs
-30 0.00% 0.00% -0.06% -0.06% 0.01% 0.01%
-29 0.28% 0.28% -0.38% -0.45% -0.19% -0.18%
-28 0.19% 0.47% 0.18% -0.27% 0.35% 0.17%
-27 0.17% 0.65% 0.53% 0.26% 0.62% 0.79%
-26 0.21% 0.86% 0.28% 0.53% -0.09% 0.70%
-25 0.07% 0.93% -0.14% 0.40% -0.15% 0.55%
-24 -1.82% -0.89% -1.90% -1.51% -1.92% -1.37%
-23 0.44% -0.45% -0.03% -1.54% 0.43% -0.95%
-22 -0.10% -0.56% 0.14% -1.40% -0.07% -1.02%
-21 -0.23% -0.79% 0.43% -0.97% 0.71% -0.31%
-20 0.57% -0.22% -0.08% -1.05% -0.03% -0.34%
-19 0.11% -0.11% -0.36% -1.42% -0.29% -0.63%
-18 -0.12% -0.23% 0.19% -1.23% 0.26% -0.37%
-17 -0.16% -0.40% -0.19% -1.42% -0.02% -0.39%
-16 -0.17% -0.56% 0.14% -1.28% -0.07% -0.46%
-15 0.48% -0.08% 0.50% -0.78% 0.30% -0.15%
-14 0.69% 0.60% 0.49% -0.29% 0.68% 0.53%
-13 -0.38% 0.23% 0.26% -0.03% -0.59% -0.06%
-12 -0.02% 0.21% 0.03% 0.00% 0.73% 0.66%
-11 0.07% 0.28% 0.10% 0.09% -0.18% 0.48%
-10 0.30% 0.58% 0.59% 0.68% 0.50% 0.98%
-9 -0.32% 0.26% -0.24% 0.44% 0.21% 1.20%
-8 0.04% 0.29% 0.17% 0.60% 0.54% 1.74%
-7 0.19% 0.48% -0.13% 0.47% -0.32% 1.41%
-6 0.47% 0.96% 0.81% 1.28% 0.12% 1.53%
-5 0.24% 1.19% 0.35% 1.63% 0.27% 1.81%
-4 0.25% 1.45% -0.56% 1.07% -0.47% 1.34%
-3 0.04% 1.48% -0.09% 0.98% 0.06% 1.40%
-2 0.79% 2.27% 0.45% 1.43% 0.42% 1.82%
-1 0.91% 3.18% -0.34% 1.09% 0.25% 2.06%
0 2.97% 6.11% 0.11% 1.21% 0.23% 2.30%
CAARs (-x , -y)
(-30, -21) -0.79% -0.97% -0.31%
(-20, -11) 1.07% 1.06% 0.79%
(-10, -1) 2.90% 1.54% 2.25%
(-30, -1) 3.18% 1.09% 2.06%
(-20, 1) 3.97% 2.18% 2.37%
(-30, 0) 6.11% 1.21% 2.30%
Announcement day AGoogle1
Announcement day AGoogle1
AGoogle2
AGoogle2

37. 36
When analysing the results recorded in Table 8, the first noticeable aspect is that both
AGoogle1 and AGoogle2 present share price run-ups, recording CAARs (-30,0) of 1.21 and
2.30%, respectively. Therefore, both measures fail to explain entirely the share price run-up of
the study sample discovered by the former study (6.11%). Although these values are smaller
than the former study, they present evidence of the existence of abnormal returns before the
market is aware of the intended mergers. The two targets’ run-ups can be appreciated with
more clarity in Figure 5. With AGoogle2 showing the strongest upward trend.
It also worth mentioning that AARs from AGoogle1 and AGoogle2 fluctuate in a more random
manner until closer to day 0 than the former study. Thus, the run-up period is reduced in length
in both measures comparing to the former study. They present more negative values during the
last 10 days of the study period, 5 and 2, respectively versus 1 negative value from the original
study. In turn, it affects the CAARs trends in both Google measures which start to show
significant values around day -6 and day -9, respectively. To mention that in the former study,
the first significant values appear around day -15.
In both Google measures, day 0 (when investors are aware of intended mergers for the first
time) present positive AARs, 0.11% for AGoogle1 and 0.23% for AGoogle2. Although lower
in significance comparing to the announcement day in the former study, they still present
evidence of the market reacting efficiently to new information. It might also be considered as
evidence for data robustness from the outlier calculations.
Finally, we divided the study period into blocks of 10 days to understand to what extent Google
Search study can explain the targets’ run-up pattern. However, the results are analysed in more
depth in the next chapter “Findings”.
Figure 6: Abnormal returns run-ups from AGoogle1/AGoogle2
The figure shows the cumulative abnormal returns recorded prior AGoogle1 and AGoogle2 yielded by the Google search
study.