Accepted Manuscript

Title: The Source of Financing in Mergers and Acquisitions

Author: id="aut0005" author-id="S1062976917300078-

0072e9fd89440a70915f7c0c2c3b6e5c"> Mario
Fischer

PII: S1062-9769(17)30007-8
DOI: http://dx.doi.org/doi:10.1016/j.qref.2017.01.003
Reference: QUAECO 999

To appear in: The Quarterly Review of Economics and Finance

Received date: 13-2-2016

Revised date: 5-11-2016
Accepted date: 4-1-2017

Please cite this article as: Mario Fischer, The Source of Financing in Mergers and
Acquisitions, <![CDATA[Quarterly Review of Economics and Finance]]> (2017),
http://dx.doi.org/10.1016/j.qref.2017.01.003

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 The Source of Financing in Mergers & Acquisitions

Mario Fischer
Department of Financial Management and Capital Markets, TUM School of Management, Technische
Universitat Munchen, Arcisstr. 21, 80333 Munich, Germany

t
Abstract

ip
I directly focus on the source of nancing in takeovers instead of the common but indirect
approximation by the payment method. By examining a sample of 610 acquisitions occur-

cr
ring between 1991 and 2009, I am able to distinguish between several dierent sources of
nancing for sizeable transactions and to additionally control for any payment eect. For

us
the initial decision if the takeover should be nanced with internal funds, the completion
time and the acquirers pre-takeover characteristics (cash level, Tobins Q, and leverage)
are crucial. When deciding the source of external funds, the acquirers pre-takeover cash
level remains important, while the targets characteristics (nationality, listing, and compet-

an
ing bids) gain importance. For acquisitions that are more credit-nanced, I nd superior
short-run performance; takeovers nanced mainly with common stock issues yield poor
announcement returns. Over the three years following an acquisition, my analysis reveals
that capital markets eciently price all information at the announcement; only takeovers
M
nanced with a common stock issue signicantly underperform in subsequent years.

Keywords: Acquisition, Financing Source, Long-Run Performance, Merger, Payment

d

Method, Short-Run Performance

te

1. Introduction
p

Even though there exists overwhelming empirical evidence suggesting that cash pay-
ment outperforms stock payment in takeovers, the economic rationale for this outperfor-
mance is still up for discussion.1 This is rather shocking, as mergers and acquisitions
ce

are among the most inuential decisions for companies and nding a value-maximizing
structure of the transaction should be crucial. So far, the (implied) academic assump-
tion is that the payment method might be a valid approximation for the involved source
Ac

of nancing (for example, Faccio and Masulis, 2005; Harford, Klasa, and Walcott, 2009;
Uysal, 2011; Karampatsas, Petmezas, and Travlos, 2014; Vermaelen and Xu, 2014). In
this context, Schlingemann (2004) states that the form of payment has been used as a
proxy or substitute for the source of nancing (p. 684), and Martynova and Renneboog

Contact: mario.scher@tum.de. This version: October 2016. JEL classications: G14, G32, G34.
The results of this paper are included in my PhD thesis at Technische Universitat Munchen. Part of the
research was conducted during stays at Harvard and Yale University.
1
The terms acquisition, merger, takeover, and transaction are used interchangeably in this study. The
outperformance of cash payment is shown by Travlos (1987), Amihud, Lev, and Travlos (1990), Brown
and Ryngaert (1991), Servaes (1991), Fuller, Netter, and Stegemoller (2002), and Faccio, McConnell, and
Stolin (2006), among others.

Page 1 of 27
(2009) argue that in previous literature the term means of payment is usually considered
as synonymous to the source of takeover nancing (p. 290). The underlying rationale is
that companies only have access to a very limited amount of cash at a given point in time.
Unless the acquirer saves free cash ows long before the actual takeover, there will not be
sucient liquid assets available for cash payment and hence, additional debt is necessary
to pay in cash. On the other hand, this strand of literature assumes that stock payment
is always equal to issuing new stocks.
Although those assumptions seem reasonable at rst glance, closer examination reveals
several shortcomings. Retained earnings as one of the major sources of nancing is not
considered at all, as cash payment is assumed to be completely nanced with debt. This

t
is an obvious oversimplication if companies have sucient internal sources for their cash

ip
payment. Furthermore, it is possible that a company uses proceeds from an equity issue
to pay in cash. Also, the payment method could inuence the source of nancing. One

cr
example is that of very limited internal cash and favored cash payment, which provokes the
need of additional external nancing. In either case, the underlying nancing decision can
tremendously aect or be driven by the means of payment in an acquisition. In contrast

us
to the vast literature on the means of payment in takeovers (for example, Martin, 1996;
Ghosh and Ruland, 1998; Ismail and Krause, 2010), the underlying nancing decision has
been overlooked in empirical studies. It is at this point that the present study contributes
to the empirical research by investigating the decision for a specic nancing structure

an
and the inuence of this structure on the acquirers performance.
This paper is the rst to investigate both short-run and long-run performance of acqui-
sitions dependent on the underlying source of nancing while simultaneously considering
the payment method used the latter is crucial considering the examples given above.
M
Furthermore, the study examines driving characteristics for the nancial decision in de-
tail. The two most closely-related investigations are Bharadwaj and Shivdasani (2003) and
Martynova and Renneboog (2009), both which explicitly focus on the source of nancing
d
instead of on the payment method. As the former only consider cash-paid deals, they
do not have any variation in the method of payment. The latter apply a hand-collected
te

European sample and focus on the nancing decision as well as the announcement re-
turns. In a rst step, I expand their results on the announcement returns to a worldwide
sample which, in particular, includes the United States. In a second step, my investiga-
p

tion provides novel insights into the long-run implications of the source of nancing a
completely overlooked area. Although Bharadwaj and Shivdasani (2003) and Martynova
ce

and Renneboog (2009) estimate the capital market response to the announcement of an
acquisition dependent on the source of nancing, they do not measure any long-run eects.
This is in sharp contrast to the extensive literature on the long-run eects of the payment
method in takeovers (for example, Loughran and Vijh, 1997; Mitchell and Staord, 2000;
Ac

Akbulut, 2013).
There are several cases in which the long-run eects of nancing might be interesting.
On the one hand, nancing an acquisition with internal cash could underperform, if one
proposes a possible empire building behavior of managers with free cash ows. On the
other hand, using credit nancing for a takeover might improve not only the initial target
selection but also the integration process, as banks can help screening in the beginning and
closely monitor the later integration progress. Altogether, my investigation in this part
not only contributes to the literature on takeovers, but also helps to explain traditional
corporate nance issues. If a takeover is seen as an investment project, I empirically show
the implications of the marginal nancing of those investment projects. For most other
investment projects, any breakdown of the underlying source of nancing is arbitrary

Page 2 of 27
for outsiders. However, during takeovers, information regarding the method of payment
and/or associated sources of nancing is sometimes released and hence, the present study
is also a suitable test for traditional capital structure theories.
The remainder of the study is organized as follows. Chapter 2 outlines previous lit-
erature on the source of nancing in takeovers. Chapter 3 establishes the hypotheses for
the empirical investigation. The data and the implemented methodology are discussed
in Chapter 4. All empirical results, including descriptive statistics of my sample, are
presented in Chapter 5. Chapter 6 concludes the present study.

2. Literature Overview

t
ip
Chronologically, the rst study directly investigating the source of nancing (instead
of the payment method) is that of Datta and Iskandar-Datta (1995).2 Their focus is

cr
on the abnormal announcement returns for stockholders and bondholders in 63 partial
acquisitions from 1982 to 1990. As a consequence of the research focus, the sample size is
rather small, as they only consider very large partial acquisitions and restrict the sample
to acquirers with outstanding liquid bonds.3 For the source of nancing, their information

us
is based on newspaper articles, and they dierentiate between the three categories of
stock, debt, and internal cash. Overall, their results indicate a loss for bondholders and
normal returns for stockholders around the announcement. Subdivided into the sources of

an
nancing, debt nancing is worse than stock nancing for bondholders. For stockholders,
the contrary holds true, with stock nancing as the worst source of nancing.
The rst of the two major direct investigations of the source of nancing in takeovers
is conducted by Bharadwaj and Shivdasani (2003) on a sample consisting of 115 cash
M
tender oers between 1990 and 1996. Bharadwaj and Shivdasani (2003) rst examine
under which circumstances bank nancing is more likely to be used. As expected, bank
nancing prevails when the acquirers cash reserve is low or the relative size of the takeover
d
is large. Overall, around 70 percent of the cash tender oers in their sample are at least
partly bank-nanced, and in half of the takeovers, bank nancing is sucient for the
te

entire transaction value. With regard to the level of corporate diversication, they do
not nd a dierence in the likelihood of bank nancing. In a second step, Bharadwaj
and Shivdasani (2003) examine the abnormal returns around the announcement of the
p

takeover. In an univariate framework, their study demonstrates that cumulative abnormal

returns for acquirers are higher when the acquisition is nanced entirely with bank debt
ce

compared to acquisitions which are fully nanced with internal funds. The former is
associated with highly signicant cumulative abnormal returns of 2.08 percent (two-day
event window) and 4.00 percent (three-day event window), whereas the latter is related
to insignicant cumulative abnormal returns of -0.32 percent (two-day event window)
Ac

and 0.54 percent (three-day event window). Those results basically also hold true in the
multivariate setting, as the cumulative abnormal returns around the announcement are
positively related to the fraction of bank nancing. The positive cumulative abnormal
returns in bank-nanced takeovers mostly occur for rms with poor performance and
high information asymmetries. Therefore, Bharadwaj and Shivdasani (2003) conclude
that their evidence is consistent with a monitoring and certication function of banks for

2
This chapter focuses on direct investigations of the source of nancing in takeovers and excludes studies
which allow conclusions based on an indirect approach (for example, Schlingemann, 2004; Harford, Klasa,
and Walcott, 2009; Elsas, Flannery, and Garnkel, 2014).
3
They argue that partial acquisitions have the advantage of allowing to exclusively investigate the
signaling eect while holding tax eects, payment method, and mood of the takeover mostly constant.

Page 3 of 27
favorable acquisitions. As their sample exclusively consists of cash tender oers, the study
is unable to dierentiate between payment and nancing eects. This gap is partly lled
by the second major direct investigation of Martynova and Renneboog (2009), who link
method of payment and the sources of nancing in takeovers.
Martynova and Renneboog (2009) hand-collect information on a sample of 1,361 Eu-
ropean takeovers between 1993 and 2001. However, as the news announcements used are
rather vague and for the most part do not disclose very detailed information, they face
some limitations. More precisely, the proportional breakdown is unavailable in news an-
nouncements and hence, the resulting categories for the source of nancing in the study of
Martynova and Renneboog (2009) are internal funds, equity issues, debt issues, and a com-

t
bination of equity and debt issues. The latter three categories may include an unspecied

ip
proportion of internal funds. Similarly to Bharadwaj and Shivdasani (2003), Martynova
and Renneboog (2009) also proceed in two steps: investigating the determinants for the

cr
source of nancing and then examining the relevance of the nancing decision on the cu-
mulative abnormal returns of the acquirers around the announcement. They show that
the nancing decision of the bidding rm is explained by the pecking-order theory, con-

us
siderations of the acquirers cost of capital, and a bidders preference for the means of
payment. In line with the pecking-order theory, internal funds are primarily used by
cash-rich rms. When internal funds are insucient, companies with low leverage prefer
borrowing, while rms with high pre-takeover stock price run-ups favor an equity issue.

an
Combining the payment method and the source of nancing, Martynova and Renneboog
(2009) show that strategic preferences for the payment method inuence the underlying
nancing decision. If an acquirer seeks risk sharing and therefore prefers stock payment,
equity nancing is applied. However, if the acquirer is susceptible for control change,
M
equity nancing is less likely. In their sample, approximately one-third of fully cash-paid
takeovers is partly nanced with external debt or equity. When investigating the cumu-
lative abnormal returns, Martynova and Renneboog (2009) conrm the earlier results of
d
Bharadwaj and Shivdasani (2003) that fully cash-paid acquisitions nanced with internal
funds signicantly underperform those nanced with debt. Whereas the former are asso-
te

ciated with 0.79 percent cumulative abnormal returns over the three-day event window,
the latter are associated with 1.32 percent. Equity nancing (independent of the means of
payment) is associated with even lower cumulative abnormal returns of 0.49 percent over
p

the three-day event window. Independent of the source of nancing, full cash payment
is related to higher cumulative abnormal returns than is full stock payment. The out-
ce

performance of debt-nanced takeovers can be explained by debt conveying the absence

of stock overvaluation and helping to limit empire building of managers. Martynova and
Renneboog (2009) conclude that the underlying nancing source has a signicant impact
on the market reaction around the announcement of a takeover.
Ac

Similar to the empirical approach in my investigation, Dittmar, Li, and Nain (2012)
and Vladimirov (2015) also use the information on the source of nancing provided by SDC
Platinum. Even though it is not the main focus of the former, the related variable on debt
nancing is positive and statistically signicant in explaining the acquirers cumulative
abnormal returns around the announcement. Their study also suggests that there is
a relation between debt nancing and cash payment, as debt might be used to pay in
cash. More recently, Vladimirov (2015) models the nancial decision in takeovers. His
study contains an empirical investigation mostly concerned with the determinants for
non-debt nancing and the implications of non-debt nancing for the takeover premium.
Nevertheless, the respective variable of non-debt nancing is negative and signicant in
explaining the acquirers cumulative abnormal returns around the announcement.

Page 4 of 27
3. Hypotheses

As the empirical investigation sheds light onto three aspects of the source of nancing
(the initial decision of how to nance a takeover, the announcement eect, and the long-run
implications), the development of hypotheses progresses in a similar way.
The rst hypothesis is concerned with the initial decision of the acquirer on when
to use which source of nancing. Based on previous research on the payment method
(which assumes a relation to the source of nancing), the proportion of cash as means
of payment should be lower when the proportion of common stock as source of nanc-
ing rises. Furthermore, several studies (for example, Amihud, Lev, and Travlos, 1990;

t
Bharadwaj and Shivdasani, 2003; Martynova and Renneboog, 2009; Chemmanur, Paeglis,

ip
and Simonyan, 2009) suggest that internal nancing and cash payment are primarily used
in smaller takeovers and that additional external nancing sources are needed for larger
takeovers. This is the case when acquirers have insucient internal resources to nance

cr
the target. By contrast, acquirers nancing the transaction value with internal funds or
paying with cash are expected to have more net cash before the takeover. Last but not
least, my investigation should strengthen previous studies with regard to completion time.

us
One reason for the use of cash payment is the faster completion of the takeover in par-
ticular in the context of a competitive bidding process (for example, Chemmanur, Paeglis,
and Simonyan, 2009; Chen, Chou, and Lee, 2011; Oenberg and Pirinsky, 2015).

an
Hypothesis 1: Several acquirer and target characteristics inuence the decision
on the source of nancing.
M
For the examination of abnormal returns in the short run and long run, several expec-
tations are based on considerations of traditional capital structure theories. The trade-o
theory suggests that the optimal level of debt balances the benets of an additional tax
shield and the drawback of higher costs of nancial distress (Kraus and Litzenberger, 1973)
d

however, this theory provides no clear indication as to which source of nancing should
be used in an acquisition. By contrast, the pecking-order theory presents an explicit rec-
te

ommendation for nancing new projects (Myers, 1984; Myers and Majluf, 1984). Dividing
the nancing sources into internal retained earnings, external debt, and external equity,
this theory assumes the presence of information asymmetry in which managers know the
p

true value of the rm and shareholders do not. In this setting, shareholders will assume
an overvaluation of their stocks when new equity is issued. Consequently, managers try
ce

to avoid issuing new equity, as they want to prevent any decrease of the share price. In-
stead, they prefer to use retained earnings in order to nance new projects, as those funds
have no adverse selection problem. If internal funds are insucient to nance the project,
Ac

managers will use external debt before issuing new equity because the signaling eects of
debt are smaller. Based on the free cash ow theory of Jensen (1986), one might expect
that available cash may be used to nance acquisitions even beyond the optimal size for
shareholders (often referred to as empire building). Besides binding these free cash ows,
loans from banks might also be benecial, as banks are supposed to initially screen the
investment projects and implement ongoing monitoring in case they grant the loan (for
example, Leland and Pyle, 1977; Diamond, 1984; James, 1987). Therefore, banks help to
reduce information asymmetries for outside investors, and their willingness to lend money
for an investment project implies a signal regarding the true quality of the project.
For the short-run abnormal returns, another prediction can be drawn from the lit-
erature on the payment method. Accordingly, cash-paid takeovers should have higher
abnormal returns around the announcement day than stock-paid takeovers. Reasons for

Page 5 of 27
this may be the absence of acquirer overvaluation, more precise estimation of the targets
value,4 or the absence of job retention purposes on behalf of the acquiring rms man-
agement. Furthermore, the pecking-order theory suggests that companies should prefer
internal funding over new issues. If internal funding is insucient for the takeover, credit
nancing is preferred based on possible bank screening in the beginning and ongoing mon-
itoring of the integration process. This reasoning is strengthened by the fact that cash
payment is often nanced with internal funds or bank loans.

Hypothesis 2: In the short run, cash payment is superior to stock payment.

For the source of nancing, new issue underperforms other types of nancing

t
in particular, credit nancing by banks.

ip
My third hypothesis covers possible long-run performance deviations over the subse-
quent three years. If the above reasoning holds true, there is still the discussion of whether

cr
or not all of the eect is priced at the announcement. Considering bank nancing, the
literature suggests a positive signal based on the initial screening as well as the ongoing

us
monitoring. To what extent the latter is correctly priced at the announcement might be
questionable. The expected short-run negative performance of nancing an acquisition
with a stock issue might turn around in the long run. When managers are aware of share
price decreases after the announcement of a common stock issue, they might choose such

an
acquisitions carefully, resulting in the possible outperformance of those acquisitions in the
long run. By contrast, stock issues are often used for stock payments that are usually
underperforming in the long run5 hence, both eects might cancel each other out. Fi-
nally, one should consider internal funds. Following the free cash ow theory, one might
M
expect an underperformance of takeovers nanced with internal cash because of possible
empire building. Besides internal funds, bank loans are similar to internal cash in terms
of unrestricted use by management. Therefore, bank loans might also be used for empire
building in some cases. All these outlined remarks would imply that the capital mar-
d

ket is unable to correctly price the takeover at the announcement. This is conterminous
te

to arguing that the capital market is not ecient in pricing new information. As most
academics expect ecient capital markets, I hypothesize that acquirers do not deviate
from the expected performance in the long run and that all information is correctly priced
p

around the announcement of the takeover.

ce

Hypothesis 3: In the long run, there is no performance deviation for any kind
of nancing or for any method of payment.

4. Data and Methodology

Ac

4.1. Sample Construction

My international sample is built on two dierent databases, namely SDC Platinum for
data on takeovers and Datastream/Worldscope for accounting data and share prices. To
be included in the initial sample of takeovers, the following criteria must be satised:
Acquirer is listed, but there are no restrictions on the listing status of the target

4
Acquirers are supposed to use stock payment to share any valuation uncertainty with the targets
shareholders (Hansen, 1987).
5
The existing literature suggests that stock payment is increasingly used in times of overvaluation and
as a consequence, stock-paid takeovers underperform cash-paid takeovers in the long run, as the market
corrects this overvaluation.

Page 6 of 27
 Announcement date is between 1991 and 2009
Acquisition is completed
Acquirer takes over 100 percent of the target
Transaction value exceeds one million U.S. Dollar
Financial acquirers are excluded, identied by the primary SIC code
Ratio of transaction value to acquirer market capitalization four weeks prior the
announcement must be above ten percent and below 120 percent6

To investigate the payment method and the source of nancing in the subsequent

t
chapter, two additional restrictions are needed:

ip
Information on the source of funds is published
Percentage of stocks, cash, and others add up to 100 percent

cr
One issue in the context of long-run abnormal returns is serial acquirers, as this might
inuence the measured returns. For short-run event studies, this problem is negligible

us
because the returns are measured over very few days. As the long-run event study considers
the subsequent three years, serial acquirers are excluded.7 After that step, 1,141 takeovers
remain of which 610 are suciently covered by Datastream/Worldscope. All amounts,
including returns, are denominated in U.S. Dollar.

Criteria
an
Table 1: Sample Restrictions
Remaining
M
Initial Sample of Takeovers 12,097
Sucient Information on Payment Method 9,363
Information Regarding Source of Financing 2,035
Unusual Financing Sources (for example, mezzanine instruments) 1,983
Other Takeover in the Year before the Announcement or in the Three Years 1,141
d

Following the Eective Date

Manual Screening and Takeover-Related Variables 930
te

Covered by Datastream/Worldscope 751

Sucient Data by Datastream/Worldscope 610
p

As Table 1 shows a massive drop in observations for the disclosure of the source of
ce

nancing (loss of almost 80 percent of the sample), it might be critical to assume that
this source of nancing is randomly disclosed and hence, the resulting sample is randomly
drawn out of the universe of takeovers. To rule out biased estimators caused by the sample
selection, I implement a Heckman (1979) two-stage selection model.
Ac

The rst-stage regression is based on all takeovers of SDC Platinum that satisfy the
initial criteria mentioned. The dependent variable in this rst-stage regression is a dummy
with one if the source of funds is disclosed for the takeover. The independent variables
are similar to those described in Table 2 with the exception of the exclusion restriction.

6
This restriction assures that the takeover is substantial and has long-run impact on the acquiring rms
performance. The threshold of ten percent is in line with the takeover literature (for example, Joehnk and
Nielsen, 1974; Moeller, Schlingemann, and Stulz, 2005; Rosen, 2006; Walker, 2000). The threshold of 120
percent is needed because huge mergers might have dierent long-run implications. See Chatterjee (2000)
or Netter, Stegemoller, and Wintoki (2011) for a separate analysis of huge takeovers.
7
More precisely, acquirers are deleted if they had either completed another takeover within one year
prior to the announcement of the considered transaction or had another acquisition announcement in the
three years following the completion of the considered transaction.

Page 7 of 27
This exclusion restriction variable is supposed to inuence the decision on the disclosure
of the source of funds in the rst stage but is not supposed to inuence the second stage.
I use a newly generated variable on the disclosure of the payment method as exclusion
restriction.8 Seeing this newly generated variable as suitable exclusion restriction is based
on two arguments. First, the payment method is often used as an approximation for the
source of nancing. Second, more information is released in some takeovers, and only very
limited information is public in others. Combining those two reasons, the disclosure of
the payment method is supposed to inuence the disclosure of the source of nancing
resulting in a suitable exclusion restriction. The resulting Inverse Mills Ration (IMR) of
this rst-stage regression is then included as a control variable in Table 7.

t
ip
4.2. Source of Financing and Control Variables
The focus of this study is on the relationship of abnormal returns and sources of -

cr
nancing while taking into account dierent means of payment. Whereas SDC Platinum
provides percentages of cash and stock payment, the information for the sources of nanc-
ing is less precise. Based on textual information, variables regarding the involved sources

us
of nancing are generated. One example for the textual information is: The transaction
was nanced by bank borrowings and internally generated funds.9
In the next step, I summarize those variables into four categories: any kind of bank
nancing (Credit), common stock issue (StockIssue), debt issue (DebtIssue), and internal

an
corporate funds (InternalFunds). All variables are set to zero if the respective source of
nancing is not involved. If the source of nancing is the only source, the value of the
variable is set to one. If more than one source of nancing is used, I scale the variable
under the assumption of equal use of all involved sources of nancing.10
M
Additionally, I generate one overall variable (Financing) to capture the nancing eect
completely. This overall variable is created in accordance with my hypothesis as well as
theoretical considerations and has internal corporate funds and debt issues as base case
d

(value of zero), new common stock issues as negative values, and bank nancing as positive
values.11
te

Also included in the regression models are several rm-specic and transaction-specic
control variables based on previous empirical ndings investigating abnormal returns.
These variables are explained in Table 2 and include: CashPayment, DierentNations, Dif-
p

ferentIndustries, FriendlyTakeover, MultipleBidders, PublicTarget, CompletionTime, Ac-

quirerLeverage, AcquirerTobinsQ, RelativeSize, TransactionValue, MarketValue, and Net-
ce

Cash.

8
More precisely, the variable distinguishes between not disclosed, partly disclosed, and fully disclosed
Ac

payment method for a takeover.

9
In the corresponding takeover, Megan Media bought M JC, announced on April 2nd , 2003.
10
As this assumption might be critical for my results in case a takeover has more than one reported
source of nancing, I rerun the regressions of Table 7 with two alternative classications. First, I change
the variable denition and introduce a dummy variable for each source of nancing. The dummy variable
is one if the respective source of nancing is used independent of the total number of sources involved.
Second, I exclude all takeovers with more than one source of nancing and therefore, circumvent the
problem of a proportional breakdown. It is worth noting here that over 60 percent of the sample is
completely nanced with one source anyway. Both alternative classications lead to similar regression
results and the same conclusions as Table 7.
11
As before, I rerun the corresponding regressions of Table 7 with a newly dened dummy variable;
internal funds and debt issue are the base case, the aforementioned dummy variable for common stock
nancing enters Financing as a negative value, and credit nancing enters Financing as a positive value.
Again, the alternative classication leads to similar results as Table 7.

Page 8 of 27
 Table 2: List of Control Variables
Variable Description
CashPayment1 Percentage of cash payment in the takeover
DierentNations1 Cross-country dummy, one if dierent acquirer and target nations
DierentIndustries1 Cross-industry dummy, one if dierent industries measured by the
rst two digits of the SIC code
FriendlyTakeover1 Dummy with one if takeover is friendly
MultipleBidders1 Dummy with one if more than one bidder is involved
PublicTarget1 Dummy for status of target, one if target is listed
CompletionTime1 Time dierence in days between date announced and date eective
AcquirerLeverage2,3 Acquirer leverage in the year before the takeover announcement

t
AcquirerTobinsQ2,3 Tobins Q of acquirer the year before the takeover announcement

ip
RelativeSize1 Transaction value divided by the acquirers market value four weeks
before the takeover
TransactionValue1 Transaction value of the takeover in million U.S. Dollar

cr
MarketValue1 Market value of the acquirer four weeks before the takeover announce-
ment in million U.S. Dollar
NetCash1,2,3 Acquirers net cash at year-end before the announcement divided by

us
the acquirers market value four weeks before the takeover
1
indicates the SDC Platinum database and 2 indicates the Datastream/Worldscope database. Further-
more, 3 shows a winsorization at the bottom and top 0.5 percent.

an
4.3. Short-Run Performance
A major part of the current study is capturing possible abnormal returns as a conse-
quence of the acquisition announcement using an event study.12 In general, this abnormal
M
return can be interpreted as the dierence between the realized return and the expected
return (in other words, the normal return) for this company without the event. The crit-
ical point is how to determine the expected, or normal, return, assuming that the event
d
did not (actually) happen. I will focus on the most popular approach in recent empirical
literature for this estimation, namely, the market model (for example, Harford, Humphery-
Jenner, and Powell, 2012; Deng, Kang, and Low, 2013; Vermaelen and Xu, 2014).13 As
te

suggested by Andrade, Mitchell, and Staord (2001) and also common in recent studies
(for example, Custodio and Metzger, 2013; Fu, Lin, and Ocer, 2013; Vladimirov, 2015),
p

I use a symmetric three-day event window and an estimation period of 150 days, ending
40 days before the event window to ensure that my estimations are not biased by any
ce

rumors (Schwert, 1996).14

Next, I introduce the aforementioned market model for the return of rm i at day t:

Rit = i + i Rmt + it (1)

Ac

with E(it ) = 0 and var(it ) = 2it . Accordingly, E(Rit ) is determined in the market
model by the following equation:

E(Rit ) = i + i Rmt (2)

12
For a more detailed explanation of the used methodology, see MacKinlay (1997).
13
Unreported tests rerun Table 7 with a constant mean return model as well as the three-factor model
of Fama and French (1993) the conclusions are completely unchanged.
14
In several unreported robustness checks, I rerun the regressions of Table 7 with dierent estimation
and event windows. The implications are unchanged and the nancing variables remain signicant with
identical signs.

Page 9 of 27
 As my sample is international, the market return Rmt is approximated by the MSCI
World (provided by Datastream). After calculating the abnormal returns ARit per rm i at
day t, I aggregate those abnormal returns in two steps: (i) over the whole event window for
one rm; (ii) and subsequently, across companies. To determine the cumulative abnormal
return CARi per rm i over the event window, beginning at t2 and ending at t3 :
t3

CARi = ARit (3)
t=t2

Next, I calculate the average cumulative abnormal return ACAR as the average over

t
the cumulative abnormal returns of all N companies in the respective sample:

ip
N
1 
ACAR = CARi (4)
N

cr
i=1

4.4. Long-Run Performance

us
Typically, the abnormal returns over the short event window are very robust with re-
gard to estimated size and statistical signicance. However, this does not hold true for
abnormal returns in the long run, and several studies discuss methodological concerns
(for example, Barber and Lyon, 1997; Fama, 1998; Mitchell and Staord, 2000). Follow-

an
ing Fama (1998) and Mitchell and Staord (2000), I apply the calendar-time portfolio
approach and use the subsequent three years after the completion of the takeover as inves-
tigated period. Initially presented by Jae (1974) and Mandelker (1974), the calendar-time
portfolio approach focuses on the portfolio return of event rms. Therefore, the perfor-
M
mance of the event rm portfolio is measured relative to an asset pricing model over time.
This portfolio of event rms is constructed every month with all event rms of the last
three years. As suggested by Loughran and Ritter (2000) and Bouwman, Fuller, and Nain
d

(2009), I use an equally-weighted portfolio.

The used asset pricing model consists of the three factors of Fama and French (1993)
te

and the momentum factor of Carhart (1997):15

Rpt Rf t = p + p (Rmt Rf t ) + sp (SM Bt ) + hp (HM Lt ) + wp (W M Lt ) + pt (5)

p

with the return of the event rm portfolio Rpt , the risk-free rate Rf t , the size factor SM Bt ,
ce

the value factor HM Lt , and the momentum factor W M Lt , all in month t. The used
standard errors for p are autocorrelation-adjusted over three months following Newey
and West (1987).
Loughran and Ritter (2000) argue in favor of a weighted least squares regression in-
Ac

stead of an ordinary least squares setting with the number of event rms in the portfolio
as monthly weight. The underlying argument is that acquisitions tend to be clustered
over time, and an ordinary least squares approach would equally weigh all monthly port-
folio observations even though there is a varying number of event rms included in every
months portfolio resulting in underweighted event rms in times of high takeover activ-
ity. Hence, I will show both alphas based on an ordinary least squares regression and a
weighted least squares regression.

15
The monthly global and U.S. factors are provided by Kenneth R. French, accessible under
http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html (May 2015).

10

Page 10 of 27
5. Empirical Results

5.1. Descriptive Statistics

To provide some information on my sample, Table 3 shows descriptive statistics of
the major variables. The rst four variables are based on the textual information per
takeover and display a preference for credit nancing. An average of 57.08 percent of
the transaction value is nanced by banks, and more than 25 percent of the takeovers
use bank loans as the sole source of nancing. By contrast, nancing a takeover with a
common stock or a debt issue is rather rare, with a low mean value of 9.21 percent and
7.47 percent, respectively. This is not surprising, considering that new issues are regarded

t
as costly. Around one-fourth of the transaction value is nanced with internally available

ip
funds in my sample. The predominance of credit nancing is also visible in the average
value for Financing of 47.87 percent.

cr
Table 3: Summary Statistics of Firm- and Takeover-Specic Variables
Variable Obs. Mean Std. Dev. 25th Pctl. 50th Pctl. 75th Pctl.

us
Credit 610 57.08% 39.53% 0.00% 50.00% 100.00%
StockIssue 610 9.21% 25.15% 0.00% 0.00% 0.00%
DebtIssue 610 7.47% 23.37% 0.00% 0.00% 0.00%
InternalFunds 610 26.24% 33.57% 0.00% 0.00% 50.00%

an
Financing 610 47.87% 54.91% 0.00% 50.00% 100.00%
CashPayment 610 79.72% 29.70% 65.17% 100.00% 100.00%
DierentNations 610 75.41% 43.10% 100.00% 100.00% 100.00%
DierentIndustries 610 37.70% 48.50% 0.00% 0.00% 100.00%
FriendlyTakeover 610 95.57% 20.58% 100.00% 100.00% 100.00%
M
MultipleBidders 610 2.46% 15.50% 0.00% 0.00% 0.00%
PublicTarget 610 35.74% 47.96% 0.00% 0.00% 100.00%
CompletionTime 610 95.92 115.58 36.00 64.00 112.00
AcquirerLeverage2 610 30.07% 26.66% 8.58% 27.43% 42.81%
d

AcquirerTobinsQ2 610 1.96 1.16 1.28 1.63 2.21

RelativeSize 610 38.12% 24.96% 17.52% 31.17% 51.19%
te

TransactionValue1 610 1,263.87 4,340.58 61.46 212.14 786.20

MarketValue1 610 3,709.22 10,605.51 202.05 729.39 2,322.63
NetCash2 610 5.66% 20.30% -0.81% 2.92% 10.51%
p

CAR 610 1.51% 9.15% -2.71% 0.76% 5.15%

1 2
indicates values in million U.S. Dollar and shows a winsorization at the bottom and top 0.5 percent.
ce

For the method of payment mostly used as an approximation of the source of nancing
in recent research I observe that takeovers are, on average, paid with a cash proportion of
79.72 percent and that the rst quartile already has over 65 percent of cash payment. The
Ac

predominance of cash payment compared to stock payment is in line with other studies
(for example, De, Fedenia, and Triantis, 1996; Goergen and Renneboog, 2004; Faccio,
McConnell, and Stolin, 2006; Akbulut, 2013). Furthermore, it underlines the high average
of credit nancing when it is assumed that cash payments are mostly nanced with bank
debt. In over half the considered takeovers, full cash payment is utilized.
Most takeovers in my sample are cross-border acquisitions, meaning that the target
and the acquiring rm are located in dierent countries. Approximately 38 percent are
diversifying takeovers with a target in a dierent industry. As in the earlier study of
Andrade, Mitchell, and Staord (2001) and in the large sample of Betton, Eckbo, and
Thorburn (2008), most takeovers are friendly and only involve one bidder. Approximately
36 percent of the targets are listed before the takeover, indicating a high proportion of
private targets again, the value is close to Betton, Eckbo, and Thorburn (2008). The

11

Page 11 of 27
average (median) time between the announcement and the completion is approximately 96
days (64 days); both values are comparable to the sample of Betton, Eckbo, and Thorburn
(2008) for public acquirers. The mean (median) pre-takeover leverage of the acquirer is
30.07 percent (27.43 percent), and the pre-takeover Tobins Q of the acquirer has a mean
(median) value of 1.96 (1.63). For both variables, the variation is within a plausible range.
For relative size, most takeovers are in a range between 20 percent and 50 percent. The
mean (median) relative size equals 38.12 percent (31.17 percent).16 The mean (median)
transaction value is 1.26 billion U.S. Dollar (212.14 million U.S. Dollar), with a very high
standard deviation. This indicates that a few very large transactions are included in the
sample and is in line with Aktas, Bodt, and Roll (2013). The same holds true for acquirer

t
size with a mean (median) value of 3.71 billion U.S. Dollar (729.39 million U.S. Dollar).

ip
One decisive characteristic of a successful takeover is a positive announcement return.
First descriptive insights show that takeovers in the sample trigger a positive market

cr
reaction at the announcement with a mean (median) cumulative abnormal return of 1.51
percent (0.76 percent). However, the variation is large, and a decent size of acquirers in
my sample experiences negative announcement returns. One-quarter of acquirers loses at

us
least 2.71 percent of their respective market capitalization in the three-day window.
Before turning to multivariate statistics, Table 4 categorizes several rm- and takeover-
specic variables separated by the source of nancing. The last ve columns show the
statistical signicances of dierences in mean values between characteristics of takeovers

an
fully nanced with one source and other takeovers.
There are two clear trends for the relative size of a takeover and the implemented source
of nancing. First, credit-nanced takeovers appear relatively larger than do takeovers
without any bank involvement. If no credit nancing is involved, the average relative size
M
is 31.92 percent. For fully credit-nanced takeovers, the average relative size increases
to 43.57 percent. The resulting dierence is highly signicant. Second, there is a clear
tendency toward smaller relative size with increasing internal funds. Not only is the
d
resulting dierence signicant again, but so is the dierence of fully internal funds to
credit nancing. I do not observe a trend for the proportions of new issue nancing, as
te

the average relative sizes for takeovers without any bond or stock nancing are close to
the relative sizes of takeovers that are fully nanced with bonds or new stocks. However,
the largest averages in relative sizes are shown for partly new issue-nanced takeovers in
p

both cases. The results suggest that rms use internal nancing for smaller takeovers and
that they need additional external sources (credit or new issue) for larger takeovers.
ce

As expected, the source of nancing is strongly related to the cash level of the acquirer.
Firms are more likely to seek bank nancing in the form of loans when their pre-merger
level of cash is low. Starting at 8.88 percent without any credit, an acquirer in a fully
credit-nanced takeover has average net cash of only 2.50 percent. The resulting dierence
Ac

in means is highly signicant. For credit nancing, the trends with regard to relative size
and net cash are both in line with Bharadwaj and Shivdasani (2003). A similar trend for
net cash can also be seen for bond nancing, whereas companies with very low cash reserves
might even use the proceeds not only for the takeover itself, but also to restructure their
capital structure and receive additional cash reserves. For internal nancing of takeovers,
I observe the expected trend of higher proportions if the rm has more cash. This increase
in net cash is very plausible and in line with the ndings of Martynova and Renneboog
(2009). The average net cash of 1.69 percent for acquirers that do not use any internal

16
Note that the minimum and maximum relative size is attributable to the sample selection criteria as
outlined in Chapter 4.

12

Page 12 of 27
 Table 4: Average Characteristics Based on the Source of Financing
Dierence of Fully to
Source of Not Not Stock- Debt- Internal-
Financing Involved Partly Fully Involved Credit Issue Issue Funds
RelativeSize
Credit 31.92% 36.76% 43.57% *** ** ***
StockIssue 38.02% 40.15% 36.65% **
DebtIssue 38.02% 42.37% 34.72% ** *
InternalFunds 41.69% 35.87% 26.62% *** *** ** *
NetCash

t
Credit 8.88% 6.72% 2.50% *** * ***

ip
StockIssue 6.01% 2.35% 5.45% ** ***
DebtIssue 6.55% 0.71% -4.70% *** * ** ***
InternalFunds 1.69% 8.16% 18.51% *** *** *** ***

cr
AcquirerLeverage
Credit 26.78% 28.05% 34.21% *** ** ***
StockIssue 29.95% 35.33% 23.80% ** **
DebtIssue 29.13% 39.32% 35.93% * ** ***

us
InternalFunds 34.36% 25.35% 22.48% *** *** ***
CompletionTime
Credit 98.52 99.36 90.88 *
StockIssue 97.00 104.25 66.38 ** * *

an
DebtIssue 94.87 108.43 99.76
InternalFunds 89.14 99.08 121.27 *
CashPayment
Credit 74.48% 85.86% 77.28% **
M
StockIssue 80.92% 79.60% 61.46% ** ** * **
DebtIssue 80.02% 75.72% 79.34% *
InternalFunds 76.50% 85.08% 79.88% **
Average values of the variables RelativeSize, NetCash, AcquirerLeverage, CompletionTime, and CashPay-
d

ment are displayed for the dierent source of nancing. *, **, and *** indicate a signicant dierence
of the mean values on a ten-, ve-, and one-percent level. The last ve columns compare the mean of
te

fully nanced takeovers to takeovers nanced without that specic source, to fully credit nancing, to fully
stock nancing, to fully bond nancing, and fully internal nancing, respectively.
p

funds implies that those acquirers do not have access to retained earnings regardless of
their preferences for the source of nancing. No trend is observed for stock nancing. It is
ce

interesting to notice that acquirers which use only internal funds have signicantly higher
net cash positions than acquirers which exclusively use another source of nancing.
Even though the variation for average leverage of the acquirer is rather low if cate-
Ac

gorized by the source of nancing, I observe two highly signicant trends for credit and
internal nancing. Whereas the former is connected to acquirers with higher pre-takeover
leverage, the opposite holds true for acquirers using internal funds. As for the relative
size, stock and bond issues behave in a similar way, with the highest leverage in cases of
partly nancing the takeover with the new issue.
The averages for completion time are more inconclusive, and no clear and signicant
trends are observed. It seems that mixed nancing is generally associated with slower
takeover completion and that more internal nancing causes slower execution. The rela-
tively long completion time for full internal nancing of 121.27 days might be related to
the acquirers high net cash for those takeovers. It seems reasonable that those acquir-
ers start saving retained earnings before the actual takeover in order to have sucient
net cash. If that is the case, those takeovers should have no time pressure and therefore,

13

Page 13 of 27
longer completion time is plausible. Interestingly, partly new issue-nanced takeovers have
also very long completion times, with an average of 104.25 days for stock nancing and
108.43 days for bond nancing. This is in line with the previous results, in which partly
new issue nancing is related to the largest relative size assuming that larger relative size
accompanies longer execution time.
Last but not least, I consider the connection between source of nancing and the actual
payment method. For credit nancing, the mean proportion of cash payment for takeovers
without any credit nancing (74.48 percent) is almost equal to the cash proportion with
full credit nancing (77.28 percent). As a consequence, the resulting dierence lacks in
statistical signicance. For stock nancing, I observe a clear and signicant trend, as the

t
proportion of cash payment decreases with a rise of stock nancing. Takeovers without

ip
a common stock issue have an average 80.92 percent cash payment, whereas takeovers
with complete stock nancing have an average of 61.46 percent cash payment. This result

cr
is reasonable because issued stocks can be used as stock payment, and therefore, the
proportion of cash payment decreases. For bond issues and internal funds, the results are
inconclusive.

us
5.2. Choice of the Source of Financing
The rst major contribution of the presented study is a close look at the decision on how
to nance a takeover. Whereas Bharadwaj and Shivdasani (2003) use single regressions,

an
which are basically not connected to each other and always explain the degree of bank
involvement, Martynova and Renneboog (2009) use a connected model to explain the
source of nancing conditioned on the payment method. Based on the above descriptive
statistics, which suggest that acquirers use internal nancing for smaller takeovers and
M
that they need additional external sources (credit or new issue) for larger ones, I proceed
with a dierent approach. I claim that the source of nancing is decided in a two-stage
process as shown in Figure 1, where internal nancing is preferred over external nancing
similar to the conclusions of the pecking-order theory. If the acquirer decides to use
d

additional external funds, the external source will be chosen in a second step. Again,
te

this is in line with the descriptive statistics, as I observe a trend of higher proportions of
internal nancing if the acquirer has more cash.
p

Figure 1: Two-Step Decision on the Source of Financing

ce

(1) Only
(2) Credit
Internal
Financing
Funds
Source of Internal
Ac

Financing Funds
External
(3) Bond
Funds
Financing
Needed

(4) Stock
Financing
The number in parenthesis corresponds to the regression number in Table 5.

To describe this structure empirically, I use a sequential probit model, estimating the
source of nancing and the payment method simultaneously. The results of this model are

14

Page 14 of 27
shown in Table 5. Regression (1) describes the rst step of choosing full internal nancing
or at least some external nancing. Regression (2), Regression (3), and Regression (4) of
Table 5 describe the decision of how much credit, bond, and stock nancing, respectively.
Note that Regression (2), Regression (3), and Regression (4) have 66 fewer observations
than Regression (1), as 66 takeovers are fully nanced with internal funds and hence, are
not included in the second step of the sequential probit model. Also note, that the probit
regressions of Table 5 estimate the source of nancing in the upper half and the proportion
of cash payment in the bottom half simultaneously.
Starting with the rst step of the decision process in Regression (1), the (expected)
completion time, acquirers pre-takeover leverage, acquirers pre-takeover Tobins Q, and

t
acquirers pre-takeover cash level signicantly explain whether or not acquirers rely solely

ip
on internal funds. As expected, lower acquirers pre-takeover leverage and higher cash
reserves increases the probability that the acquirer uses only internal funds. Even though

cr
the actual completion time of the takeover can only be measured afterwards, longer com-
pletion time is associated with a higher probability of full nancing with internal funds.
This could be explained by the fact that, on average, the realized completion time provides

us
an estimate of the expected completion time before the takeover. Hence, if the manage-
ment wants to close the deal relatively quickly and not wait for a sucient accumulation
of free cash ows, they tend to use some sort of external funds.
This changes for the second step in Regression (2), Regression (3), and Regression (4)

an
though. Besides the acquirers pre-takeover level of cash, all other acquirer characteristics
lose their signicance. Instead, characteristics of the takeover and the involved target
gain importance, namely the dummy for dierent nations, the listing status of the target,
and the dummy for multiple bidders. Low levels of acquirers cash make it unlikely that
M
the takeover will be nanced by bonds. If a target in another country is acquired, bank
nancing is likelier and stock nancing unlikely. The latter is also rarely used for public
targets. As expected, bond (stock) nancing is preferred (prevented) if multiple bidders
d
are involved. The rationale might be that cash payment is needed in those cases and
hence, issuing bonds is more suitable than issuing stocks.
te

For the simultaneously estimated proportion of cash payment, the results seem con-
stant over both steps. Whereas listed targets and targets in dierent countries are less
likely to receive cash payment, takeovers with multiple bidders mostly involve cash pay-
p

ment. Even though the estimated coecients for the cash payment regressions are very
stable over the two stages, indicates that all but Regression (3) have a signicant con-
ce

nection to the upper half (estimating the source of nancing).

Overall, Hypothesis 1 is mostly conrmed. Whereas acquirer characteristics are sig-
nicant in the rst step, target and takeover characteristics inuence the second step of
the nancing decision as well as the method of payment.
Ac

5.3. Short-Run Abnormal Returns

Table 6 presents the three-day average cumulative abnormal returns around the initial
announcement of the takeover, again classied by the involvement of the dierent sources
of nancing. The last column shows the dierences in mean values between takeovers
fully nanced with one source of nancing and takeovers nanced without that particular
source of nancing.
The results suggest clear tendencies for credit and new issue nancing but not for in-
ternal funds. Whereas my sample of takeovers without any credit nancing has abnormal
returns that are not statistically dierent from zero, fully credit-nanced acquisitions re-
alize the highest abnormal returns, with 2.15 percent. Not only is the latter value highly

15

Page 15 of 27
 Table 5: Sequential Probit Regressions for Simultaneously Choosing Financing and Payment
(1) (2) (3) (4)
First Step Second Step
Dependent Variable InternalFunds Credit DebtIssue StockIssue
RelativeSize -0.125 0.592 -0.137 0.153
(0.71) (0.85) (0.91) (0.90)
ln(CompletionTime) 0.157** 0.0828 -0.0450 0.0479
(0.06) (0.08) (0.09) (0.07)
ln(TransactionValue) -0.389 -0.149 0.0834 -0.0705
(0.27) (0.32) (0.36) (0.33)

t
AcquirerLeverage -0.632*** 0.0637 0.161 0.314

ip
(0.23) (0.25) (0.28) (0.24)
AcquirerTobinsQ -0.102** -0.0667 0.0165 0.0813
(0.05) (0.05) (0.06) (0.05)

cr
NetCash 1.366*** 0.698* -1.224*** -0.242
(0.34) (0.37) (0.46) (0.36)
ln(MarketValue) 0.361 0.172 0.0749 -0.0721
(0.27) (0.33) (0.36) (0.34)

us
DierentNations -0.0438 0.288** 0.209 -0.584***
(0.13) (0.15) (0.19) (0.15)
DierentIndustries 0.0369 -0.109 -0.0531 -0.0524
(0.11) (0.14) (0.15) (0.14)

an
PublicTarget 0.141 0.129 -0.101 -0.379**
(0.13) (0.17) (0.19) (0.17)
MultipleBidders -0.144 0.286 0.799** -5.026***
(0.34) (0.52) (0.38) (0.25)
M
Constant -0.748 -0.0242 -2.125** -0.105
(0.64) (0.79) (0.88) (0.85)
Dependent Variable CashPayment
RelativeSize 0.314 0.657 0.652 0.603
d

(1.12) (1.24) (1.20) (1.22)

ln(CompletionTime) -0.0307 -0.0589 -0.0691 -0.0611
te

(0.15) (0.19) (0.19) (0.19)

ln(TransactionValue) -0.156 -0.324 -0.283 -0.244
(0.47) (0.52) (0.50) (0.52)
AcquirerLeverage -0.360 -0.230 -0.232 -0.285
p

(0.28) (0.26) (0.26) (0.26)

AcquirerTobinsQ -0.0801 -0.113 -0.0916 -0.0957
ce

(0.08) (0.07) (0.08) (0.08)

NetCash 1.130* 0.373 0.324 0.341
(0.59) (0.41) (0.41) (0.41)
ln(MarketValue) 0.305 0.498 0.474 0.427
Ac

(0.49) (0.53) (0.52) (0.53)

DierentNations -0.640** -0.869** -1.001*** -0.979***
(0.27) (0.36) (0.37) (0.37)
DierentIndustries 0.286 0.402* 0.358 0.373
(0.20) (0.24) (0.24) (0.25)
PublicTarget -0.796*** -0.976*** -1.003*** -1.015***
(0.21) (0.23) (0.23) (0.23)
MultipleBidders 5.251*** 5.494*** 5.799*** 5.196***
(0.52) (0.60) (0.53) (0.52)
Constant 1.574 1.518 1.597 1.688
(1.17) (1.18) (1.25) (1.23)
0.342** 0.543*** -0.112 -0.293**
Observations 610 544 544 544
Heteroscedasticity-consistent standard errors are in parentheses. *, **, and *** indicate a signicant
dierence from zero on a ten-, ve-, and one-percent level. The model is shown in Figure 1.
16

Page 16 of 27
 Table 6: Average Cumulative Abnormal Returns
Source of Financing Not Involved Partly Fully Fully-Not Involved
Credit 0.47% 1.55%*** 2.15%*** 1.68%**
StockIssue 1.82%*** -0.02% -0.92% -2.74%*
DebtIssue 1.61%*** 1.19% -0.05% -1.66%
InternalFunds 1.41%*** 1.77%*** 1.22%* -0.19%
This table shows average cumulative abnormal returns based on the market model. *, **, and *** show a
signicance of average cumulative abnormal returns at a ten-, ve-, and one-percent level.

signicant, but so is the dierence in mean values of 1.68 percentage points between full

t
credit and no credit nancing. The preliminary conclusion is that the bank involvement

ip
increases abnormal returns. This is most likely caused by a banks monitoring and screen-
ing functions. Both help to reduce the information asymmetry between managers and

cr
shareholders concerning the takeover. If banks actually screen the acquisition decision in
detail and only provide loans in good cases, credit-nanced takeovers will convey a positive
signal to the capital market. Furthermore, shareholders might expect the bank to monitor

us
the acquirer after the takeover. A third rationale for the positive eects of credit nancing
is based on empire building considerations. Bank loans prevent managers from wasting
free cash ows for their personal benets because they limit free cash ows by additional
interest and future repayment.

an
For common stock nancing, Table 6 reveals opposing results, as average cumulative
abnormal returns decrease when the proportion of stock nancing increases. An expla-
nation for this could be an expected overvaluation by shareholders as suggested by the
M
aforementioned pecking-order theory. Therefore, capital markets react with a downward
adjustment of the share price, regardless of the actual quality of the takeover. Another
possible explanation is that an additional equity issue increases rm size. If managers
want to extend their power, this source of nancing is strictly preferred. In both cases,
d

those low abnormal returns can be interpreted as indirect costs for shareholders when the
source of nancing is a new issue. Almost the same results are obtained for bond nanc-
te

ing. However, the category of fully bond-nanced takeovers shows only marginal negative
abnormal returns.
The pecking-order theory also suggests that companies should prefer internal funding
p

over new issues. This is in line with the results in Table 6, as the abnormal returns are
higher for internal nancing than for new issue nancing. Nevertheless, no clear trend for
ce

internal nancing is visible, and the resulting dierence in mean values of -0.19 percentage
points lacks signicance.
In pairwise comparisons of the means and medians between the dierent groups that
Ac

are fully nanced with one source, only the dierences between fully credit-nanced and
fully stock-nanced takeovers are signicant. One drawback of this univariate analysis
could be that the well-known payment eect is causing these abnormal returns instead
of the actual nancing eect. Hence, Table 7 presents the results of the multivariate
regression setting. The Financing variable is dened according to the results of Table 6
with a base case of internal and debt nancing, negative values for common stock nancing,
and positive values for credit nancing.
Regression (1) of Table 7 only considers the inuence of credit nancing on cumulative
abnormal returns. As expected based on the previous univariate setting and the outlined
theory, credit nancing has a positive and signicant coecient estimate of 0.0182, even
after controlling for the payment method. This implies that a change from no credit -
nancing to full credit nancing yields 1.82 percentage points higher abnormal returns

17

Page 17 of 27
 Table 7: Regressions for Short-Run Cumulative Abnormal Returns
(1) (2) (3) (4) (5)
Dependent Variable Cumulative Abnormal Returns
Credit 0.0182**
(2.16)
StockIssue -0.0403***
(-2.83)

t
DebtIssue -0.0102

ip
(-0.92)
Financing 0.0178*** 0.0150**
(2.74) (2.02)

cr
CashPayment 0.0315** 0.0279** 0.0335*** 0.0291** 0.0303**
(2.57) (2.34) (2.75) (2.39) (2.21)
RelativeSize -0.0203 -0.0241 -0.0211 -0.0221 -0.0414

us
(-0.45) (-0.54) (-0.46) (-0.49) (-0.83)
ln(CompletionTime) 0.0000435 -0.000520 -0.000183 -0.000221 -0.000740
(0.01) (-0.09) (-0.03) (-0.04) (-0.10)
AcquirerLeverage 0.00430 0.00573 0.00531 0.00467 -0.00306

an
(0.38) (0.51) (0.47) (0.42) (-0.22)
AcquirerTobinsQ 0.00239 0.00310 0.00250 0.00267 0.00493
(0.64) (0.83) (0.67) (0.71) (1.21)
ln(MarketValue) -0.0149 -0.0189 -0.0166 -0.0163 -0.0313
(-0.77) (-0.96) (-0.84) (-0.84) (-1.32)
M
ln(TransactionValue) 0.0103 0.0139 0.0122 0.0116 0.0287
(0.47) (0.63) (0.55) (0.53) (1.08)
NetCash -0.00324 -0.00964 -0.00898 -0.00411 -0.0168
(-0.18) (-0.55) (-0.50) (-0.23) (-0.74)
d

DierentNations -0.00155 -0.00317 0.000723 -0.00330 -0.0235

(-0.16) (-0.33) (0.08) (-0.34) (-1.47)
te

DierentIndustries 0.0117* 0.0121* 0.0115* 0.0120* 0.00521

(1.68) (1.74) (1.65) (1.72) (0.65)
FriendlyTakeover -0.0125 -0.0101 -0.0113 -0.0125 -0.0448
p

(-0.23) (-0.19) (-0.21) (-0.23) (-0.79)

MultipleBidders 0.0112 0.0103 0.0114 0.0112 0.0227
(0.77) (0.72) (0.78) (0.78) (1.07)
ce

PublicTarget -0.0248*** -0.0269*** -0.0244*** -0.0261*** -0.0270**

(-2.70) (-2.86) (-2.62) (-2.82) (-2.40)
InverseMillsRatio 0.00864 0.00786 0.00413 0.0120 0.0794
(0.12) (0.10) (0.05) (0.16) (0.84)
Ac

Constant 0.0303 0.0567 0.0440 0.0348 0.327***

(0.40) (0.75) (0.59) (0.47) (2.61)
Fixed Eects No No No No Yes
Observations 610 610 610 610 610
Adjusted R2 0.039 0.044 0.034 0.044 0.064
Heteroscedasticity-consistent standard errors are in parentheses. *, **, and *** indicate a signicant
dierence from zero on a ten-, ve-, and one-percent level. Fixed eects include eective year xed eects,
industry xed eects for acquirer and target (based on the rst two digits of the respective primary SIC
code), and nation xed eects for acquirer and target. Cumulative abnormal returns are calculated with
the market model.

18

Page 18 of 27
all else being equal. The estimate is very close to the 1.68 percentage point dierence
in mean values as reported in Table 6. Besides the variable for the source of nancing,
the payment variable and the dummy for public target rms are signicant. The former
is in line with previous research on the payment eect with the expected sign and an
economically reasonable size of the estimated coecient. Fully cash-paid takeovers gen-
erate approximately three percentage points higher abnormal returns than acquisitions
without any cash payment all else being equal. The dummy for public targets has a
negative coecient estimate, meaning that public targets underperform private targets or
subsidiaries by over two percent again, all else being equal.
Similar results can be seen in Regression (2) of Table 7, whereas the sign of the stock

t
nancing variable reverses. As already anticipated in Table 6, common stock nancing

ip
has a negative inuence on announcement returns. The negative and signicant coecient
estimate indicates that a fully stock-nanced takeover underperforms by four percentage

cr
points compared to a takeover without any common stock nancing all else being equal.
Even though the variable for bond nancing is negative in Regression (3), it lacks in
statistical signicance.

us
So far, I have only considered one source of nancing simultaneously in the regressions.
This means, for example, that the eect of credit nancing in Regression (1) is compared
to all other takeovers without that particular source of nancing. However, it might be
interesting to simultaneously examine the eects of all four sources of nancing. Regression

an
(4) and Regression (5) show the corresponding regressions with the Financing variable,
with and without xed eects.
In Regression (4) of Table 7, the previous results are conrmed with a positive and
signicant estimate of almost two percentage points for the overall nancing variable.
M
This implies a two-percentage point gain (loss) when switching from full internal or bond
nancing to full credit (full stock) nancing. Moreover, a fully credit-nanced acquisition
outperforms a fully stock-nanced acquisition by over three percentage points over the
d
three-day window all else being equal. This eect is huge in economic terms, with an
average acquirers market capitalization four weeks before the takeover announcement of
te

3.71 billion U.S. dollars. Hence, the over three-percentage point change, based on the
source of nancing, leads to a 132 million U.S. Dollar wealth implication for the average
acquirers shareholders over three days.
p

Including several xed eects in Regression (5) of Table 7 does not change that result.
The coecient estimate slightly decreases to 0.0150 but keeps its signicance. Overall,
ce

the nancing eect is around half the size of the payment eect. Both are signicant,
at least at a ve-percent level throughout both regressions. Therefore, all multivariate
results conrm the preliminary conclusions based on the univariate setting in Table 6.
For robustness, I rerun the regressions with several restrictions in unreported tests. More
Ac

precisely, I only consider takeovers announced after the year 2000 (to conclude if the
nancing eect is still existing), focus on major transactions with a relative size of at
least 30 percent, and exclude private targets. Over all three regression settings, Financing
retains its expected coecient estimate and its signicance. As I observe relatively high
estimated coecients, the source of nancing has a greater inuence for more recent or
larger takeovers and if a public target is involved all three underline the importance of
this investigation for practitioners as well as academics.
Summarizing my results for the announcement eect, stock nancing underperforms
other sources of nancing, and bank-nanced takeovers perform superiorly in the short
run. Furthermore, the estimated coecients and the signicance levels are very stable.
Besides the source of nancing, both the method of payment and a possible listing of the

19

Page 19 of 27
target have steady inuences on the cumulative abnormal returns. Hypothesis 2 is almost
fully conrmed, as only the underperformance of bond nancing is not signicant in a
statistical manner.

5.4. Long-Run Abnormal Returns

After analyzing the eects of the source of nancing on the announcement returns, the
third focus of this study is on long-run implications. Based on theoretical considerations,
acquisitions nanced with internal funds might underperform other types of nancing,
as these takeovers could be conducted more with the intention of empire building than
based on their economic value. If the above reasoning regarding credit nancing holds

t
true, one might expect continued superior performance of takeovers that are at least

ip
partly bank nanced. This might be attributed to bank monitoring in the period after
the takeover or to the initial screening process before the bank loan was granted. The

cr
negative performance of nancing an acquisition with a new stock issue might continue in
the long run. Based on the considerations of using stocks as currency, one might expect
acquirers that nance their takeovers completely with a stock issue to exploit their own

us
mispricing in their favor. However, the capital market will revise this mispricing over the
subsequent years, which will result in an underperformance. Table 8 sheds light on these
considerations.

an
Table 8: Long-Run Alphas for the Calendar-Time Portfolio Approach
Portfolio Restriction Worldwide U.S.
OLS Alpha WLS Alpha OLS Alpha WLS Alpha
Credit -0.07% 0.01% -0.36% -0.10%
M
(266) (266) (265) (265)
StockIssue -0.90%** -0.89%** -2.58%*** -2.59%***
(206) (206) (165) (165)
DebtIssue -0.35% -0.77% -0.90% -0.99%
d

(255) (255) (240) (240)

InternalFunds -0.27% -0.26% -0.57% -0.79%*
te

(247) (247) (246) (246)

CashPayment -0.47% 0.00% -0.59%* -0.28%
(271) (271) (269) (269)
p

StockPayment 0.07% -0.61% 1.60% 1.67%

(189) (189) (170) (170)
ce

Alphas for the calendar-time portfolio approach using ordinary least squares (OLS) and weighted least
squares (WLS) are shown. The number of considered months are given in parentheses. StockPayment
is dened in a similar way to CashPayment. All portfolios only consider acquirers that fully nance
their takeover with one source of nancing or only use one payment method. *, **, and *** indicate a
Ac

signicant dierence from zero on a ten-, ve-, and one-percent level. Heteroscedasticity-consistent and
autocorrelation-adjusted (up to three months) standard errors are implemented.

The regular calendar-time portfolio approach with an ordinary least squares setting
presents consistent results with those obtained using weighted least squares. For credit,
bond, and internal nancing in the worldwide sample, I observe portfolio returns which are
not statistically dierent from the expectations of the Carhart (1997) four-factor model.
However, the portfolio of stock-nanced takeovers underperforms the Carhart (1997) four-
factor model with almost one percent per month. This is in line with studies showing an
underperformance after issuing equity (for example, Ritter, 1991; Loughran and Ritter,
1995; Spiess and Aeck-Graves, 1995; Butler et al., 2011). Interestingly, this does not hold
true for the portfolio of stock-paid takeovers. It should be noted that the economic eect of

20

Page 20 of 27
the underperformance of bond nancing and stock payment in the weighted least squares
setting is quite large even though it is not statistically signicant. The underperformance
of -0.77 percent (for bond nancing) and -0.61 percent (for stock payment) result in a
yearly underperformance of nine percent and seven percent, respectively, after controlling
for the overall market movements as well as size, value, and momentum eects.
In a global setting, previous studies argue that the factor calculation should be as
precise as possible for the tested sample, and one overall global model might be inap-
propriate (for example, Grin, 2002; Hou, Karolyi, and Kho, 2011; Fama and French,
2012; Hanauer and Linhart, 2015). The last two columns focus on the United States
and apply the respective factors. The interpretations are mostly similar to the worldwide

t
sample with global factors. Especially for stock nancing, the coecient estimates and

ip
signicance levels are more distinct. Interestingly, the portfolio of acquirers in stock-paid
takeovers yields economically large positive coecient estimates, even though those are

cr
statistically insignicant.
Overall, the long-run performance analysis reveals two interesting insights. First, for
credit, bond, and internal nancing, the capital market correctly prices possible implica-

us
tions of the takeover in the announcement returns. This leads to a long-run performance
which is in line with the used asset pricing model of Carhart (1997) and conrms Hypoth-
esis 3. Second, for stock nancing, an underperformance over the subsequent three years
exists, and the announcement returns have not priced in all implications contradicting

an
Hypothesis 3.

6. Conclusion
M
The presented empirical investigation contributes to the scarce literature on the source
of nancing in takeovers. So far, only Bharadwaj and Shivdasani (2003) and Martynova
and Renneboog (2009) have focused on this topic in a comparable way. However, a com-
d
prehensive strand of literature on the means of payment in takeovers exists. Connecting
both, only Martynova and Renneboog (2009) analyze the source of nancing and simul-
te

taneously consider the payment method. My study extends those investigations at least
in three dimensions. First, it includes the United States and exploits a worldwide sample.
Second, I use a two-step model to approximate for the decision process with regard to the
p

source of nancing. Third, my study examines the short-run as well as long-run eects of
the source of nancing.
ce

When looking at the initial decision on how to nance the acquisition, characteris-
tics of acquirers mainly determine the decision whether or not to use additional external
nancing. In the second step when deciding on the source of external nancing the
characteristics of targets gain importance whereas the cash level partly remains signicant.
Ac

In case of competing bids, common stock nancing seems to be less attractive.

For short-run abnormal returns, companies have the best announcement returns for
bank-nanced takeovers. Financing acquisitions with a stock issue underperforms other
sources of nancing. In line with previous research, cash-paid takeovers have better returns
on the announcement day than do stock-paid acquisitions. Economically, the payment
eect is 2.91 percent for full cash payment in comparison to full stock payment, and the
nancing eect is 3.56 percent in size if a company switches from full stock nancing of
a takeover to a complete credit-nanced takeover all else being equal for both eects.17

17
The estimates are based on Regression (4) of Table 7.

21

Page 21 of 27
Overall, the economic eect and the statistical signicance of both the payment method
and the source of nancing works similarly in the short run.
This picture is less pronounced in long-run abnormal returns than in the short run,
as even relatively large coecient estimates lack in statistical signicance. My analysis
supports the view that the capital market accurately prices the inuence of the takeover at
the announcement. One exception thereof is common stock nancing, which signicantly
underperforms the Carhart (1997) four-factor model in the subsequent three years. For
credit, bond, and internal nancing, I do not observe a systematic performance deviation.
Even though the study sheds light on dierent aspects of nancing takeovers, it faces
some limitations similar to those of Martynova and Renneboog (2009). The available

t
information on the source of nancing is rather vague. I construct my variables based

ip
on textual information on the source of nancing; a detailed breakdown per takeover is
unavailable. Although the short-run results seem to be stable with regard to that problem

cr
and the corresponding assumption, the vague information prevents several interesting
tests. For instance, it would be interesting to look at the predominant source of nancing
instead of all sources.18 For the long-run results, my study is confronted with the usual

us
problem of measuring long-run abnormal returns, and in consequence, the long-run results
may be less distinct and conclusive. Furthermore, it is critical to state that there might
be some endogeneity. Neither the takeover decision itself nor the target nor the source of
nancing are exogenous. Therefore, it is possible that certain characteristics of takeovers

an
with regard to the source of nancing or the means of payment do not underperform
themselves but are endogenously given for worse transactions.
M
Acknowledgement

I am very grateful to Andrea Schiralli for extensive proofreading and to an anonymous

reviewer for several valuable comments and suggestions.
d
p te
ce
Ac

18
Note that the calendar time-portfolio approach in Table 8 uses portfolios where the acquirer has
only one source of nancing. Therefore, this informational problem is non-existent for the calendar-time
portfolio approach.

22

Page 22 of 27
References

Akbulut, M. E. (2013). Do Overvaluation-Driven Stock Acquisitions Really Benet Ac-

quirer Shareholders? Journal of Financial and Quantitative Analysis 48.4, pp. 1025
1055.
Aktas, N., E. de Bodt, and R. Roll (2013). Learning from Repetitive Acquisitions: Evi-
dence from the Time Between Deals. Journal of Financial Economics 108.1, pp. 99
117.
Amihud, Y., B. Lev, and N. G. Travlos (1990). Corporate Control and the Choice of
Investment Financing: The Case of Corporate Acquisitions. Journal of Finance 45.2,

t
pp. 603616.

ip
Andrade, G., M. Mitchell, and E. Staord (2001). New Evidence and Perspectives on
Mergers. Journal of Economic Perspectives 15.2, pp. 103120.
Barber, B. M. and J. D. Lyon (1997). Detecting Long-Run Abnormal Stock Returns: The

cr
Empirical Power and Specication of Test Statistics. Journal of Financial Economics
43.3, pp. 341372.
Betton, S., B. E. Eckbo, and K. S. Thorburn (2008). Corporate Takeovers. Handbook of

us
Empirical Corporate Finance. Ed. by B. E. Eckbo. Vol. 2. Chap. 15, pp. 289427.
Bharadwaj, A. and A. Shivdasani (2003). Valuation Eects of Bank Financing in Acqui-
sitions. Journal of Financial Economics 67.1, pp. 113148.

an
Bouwman, C. H. S., K. Fuller, and A. S. Nain (2009). Market Valuation and Acquisition
Quality: Empirical Evidence. Review of Financial Studies 22.2, pp. 633679.
Brown, D. T. and M. D. Ryngaert (1991). The Mode of Acquisition in Takeovers: Taxes
and Asymmetric Information. Journal of Finance 46.2, pp. 653669.
M
Butler, A. W. et al. (2011). Corporate Financing Decisions, Managerial Market Timing,
and Real Investment. Journal of Financial Economics 101.3, pp. 666683.
Carhart, M. M. (1997). On Persistence in Mutual Fund Performance. Journal of Finance
52.1, pp. 5782.
d

Chatterjee, R. A. (2000). The Financial Performance of Companies Acquiring Very Large

Takeover Targets. Applied Financial Economics 10.2, pp. 185191.
te

Chemmanur, T. J., I. Paeglis, and K. Simonyan (2009). The Medium of Exchange in

Acquisitions: Does the Private Information of Both Acquirer and Target Matter?
p

Journal of Corporate Finance 15.5, pp. 523542.

Chen, S.-S., R. K. Chou, and Y.-C. Lee (2011). Bidders Strategic Timing of Acquisition
ce

Announcements and the Eects of Payment Method on Target Returns and Competing
Bids. Journal of Banking and Finance 35.9, pp. 22312244.
Custodio, C. and D. Metzger (2013). How Do CEOs Matter? The Eect of Industry
Expertise on Acquisition Returns. Review of Financial Studies 26.8, pp. 20082047.
Ac

Datta, S. and M. E. Iskandar-Datta (1995). Corporate Partial Acquisitions, Total Firm

Valuation and the Eect of Financing Method. Journal of Banking and Finance 19.1,
pp. 97115.
De, S., M. Fedenia, and A. J. Triantis (1996). Eects of Competition on Bidder Returns.
Journal of Corporate Finance 2.3, pp. 261282.
Deng, X., J.-k. Kang, and B. S. Low (2013). Corporate Social Responsibility and Stake-
holder Value Maximization: Evidence from Mergers. Journal of Financial Economics
110.1, pp. 87109.
Diamond, D. W. (1984). Financial Intermediation and Delegated Monitoring. Review of
Economic Studies 51.166, pp. 393414.

23

Page 23 of 27
Dittmar, A., D. Li, and A. Nain (2012). It Pays to Follow the Leader: Acquiring Tar-
gets Picked by Private Equity. Journal of Financial and Quantitative Analysis 47.5,
pp. 901931.
Elsas, R., M. J. Flannery, and J. A. Garnkel (2014). Financing Major Investments:
Information about Capital Structure Decisions. Review of Finance 18.4, pp. 1341
1386.
Faccio, M. and R. W. Masulis (2005). The Choice of Payment Method in European
Mergers and Acquisitions. Journal of Finance 60.11, pp. 13451388.
Faccio, M., J. J. McConnell, and D. Stolin (2006). Returns to Acquirers of Listed and
Unlisted Targets. Journal of Financial and Quantitative Analysis 41.1, pp. 197220.

t
Fama, E. F. (1998). Market Eciency, Long-Term Returns, and Behavioral Finance.

ip
Journal of Financial Economics 49.3, pp. 283306.
Fama, E. F. and K. R. French (1993). Common Risk Factors in the Returns on Stocks

cr
and Bonds. Journal of Financial Economics 33.1, pp. 356.
(2012). Size, Value, and Momentum in International Stock Returns. Journal of Fi-
nancial Economics 105.3, pp. 457472.

us
Fu, F., L. Lin, and M. S. Ocer (2013). Acquisitions Driven by Stock Overvaluation: Are
They Good Deals? Journal of Financial Economics 109.1, pp. 2439.
Fuller, K., J. Netter, and M. Stegemoller (2002). What Do Returns to Acquiring Firms
Tell Us? Evidence from Firms That Make Many Acquisitions. Journal of Finance

an
57.4, pp. 17631793.
Ghosh, A. and W. Ruland (1998). Managerial Ownership, the Method of Payment for
Acquisitions, and Executive Job Retention. Journal of Finance 53.2, pp. 785798.
Goergen, M. and L. Renneboog (2004). Shareholder Wealth Eects of European Domestic
M
and Cross-border Takeover Bids. European Financial Management 10.1, pp. 945.
Grin, J. M. (2002). Are the Fama and French Factors Global or Country Specic?
Review of Financial Studies 15.3, pp. 783803.
d
Hanauer, M. X. and M. Linhart (2015). Size, Value, and Momentum in Emerging Market
Stock Returns: Integrated or Segmented Pricing? Asia-Pacic Journal of Financial
te

Studies 44.2, pp. 175214.

Hansen, R. G. (1987). A Theory for the Choice of Exchange Medium in Mergers and
Acquisitions. Journal of Business 60.1, pp. 7595.
p

Harford, J., M. Humphery-Jenner, and R. Powell (2012). The Sources of Value Destruc-
tion in Acquisitions by Entrenched Managers. Journal of Financial Economics 106.2,
ce

pp. 247261.
Harford, J., S. Klasa, and N. Walcott (2009). Do Firms Have Leverage Targets? Evidence
from Acquisitions. Journal of Financial Economics 93.1, pp. 114.
Heckman, J. J. (1979). Sample Selection Bias as a Specication Error. Econometrica
Ac

47.1, pp. 153161.

Hou, K., G. A. Karolyi, and B.-C. Kho (2011). What Factors Drive Global Stock Re-
turns? Review of Financial Studies 24.8, pp. 25272574.
Ismail, A. and A. Krause (2010). Determinants of the Method of Payment in Mergers
and Acquisitions. Quarterly Review of Economics and Finance 50.4, pp. 471484.
Jae, J. F. (1974). Special Information and Insider Trading. Journal of Business 47.3,
pp. 410428.
James, C. (1987). Some Evidence on the Uniqueness of Bank Loans. Journal of Finance
19.2, pp. 217235.
Jensen, M. C. (1986). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.
American Economic Review 76.2, pp. 323329.

24

Page 24 of 27
Joehnk, M. D. and J. F. Nielsen (1974). The Eects of Conglomerate Merger Activity on
Systematic Risk. Journal of Financial and Quantitative Analysis 9.2, pp. 215225.
Karampatsas, N., D. Petmezas, and N. G. Travlos (2014). Credit Ratings and the Choice
of Payment Method in Mergers and Acquisitions. Journal of Corporate Finance 25,
pp. 474493.
Kraus, A. and R. H. Litzenberger (1973). A State-Preference Model of Optimal Financial
Leverage. Journal of Finance 28.4, pp. 911922.
Leland, H. E. and D. H. Pyle (1977). Informational Asymmetries, Financial Structure,
and Financial Intermediation. Journal of Finance 32.2, pp. 371387.
Loughran, T. and J. R. Ritter (1995). The New Issues Puzzle. Journal of Finance 50.1,

t
pp. 2351.

ip
(2000). Uniformly Least Powerful Tests of Market Eciency. Journal of Financial
Economics 55.3, pp. 361389.

cr
Loughran, T. and A. M. Vijh (1997). Do Long-Term Shareholders Benet From Corporate
Acquisitions? Journal of Finance 52.5, pp. 17651790.
MacKinlay, A. C. (1997). Event Studies in Economics and Finance. Journal of Economic

us
Literature 35.1, pp. 1339.
Mandelker, G. (1974). Risk and Return: The Case of Merging Firms. Journal of Finan-
cial Economics 1.4, pp. 303335.
Martin, K. J. (1996). The Method of Payment in Corporate Acquisitions, Investment

an
Opportunities, and Management Ownership. Journal of Finance 51.4, pp. 12271247.
Martynova, M. and L. Renneboog (2009). What Determines the Financing Decision in
Corporate Takeovers: Cost of Capital, Agency Problems, or the Means of Payment?
Journal of Corporate Finance 15.3, pp. 290315.
M
Mitchell, M. L. and E. Staord (2000). Managerial Decisions and Long-Term Stock Price
Performance. Journal of Business 73.3, pp. 287329.
Moeller, S. B., F. P. Schlingemann, and R. M. Stulz (2005). Wealth Destruction on
d
a Massive Scale? A Study of Acquiring-Firm Returns in the Recent Merger Wave.
Journal of Finance 60.2, pp. 757782.
te

Myers, S. C. (1984). The Capital Structure Puzzle. Journal of Finance 39.3, pp. 574
592.
Myers, S. C. and N. S. Majluf (1984). Corporate Financing and Investment Decisions
p

when Firms have Information that Investors do not have. Journal of Financial Eco-
nomics 13.2, pp. 187221.
ce

Netter, J., M. Stegemoller, and M. B. Wintoki (2011). Implications of Data Screens on

Merger and Acquisition Analysis: A Large Sample Study of Mergers and Acquisitions
from 1992 to 2009. Review of Financial Studies 24.7, pp. 23162357.
Newey, W. K. and K. D. West (1987). A Simple, Positive Semi-Denite, Heteroskedastic-
Ac

ity and Autocorrelation Consistent Covariance Matrix. Econometrica 55.3, pp. 703
708.
Oenberg, D. and C. Pirinsky (2015). How Do Acquirers Choose Between Mergers and
Tender Oers? Journal of Financial Economics 116.2, pp. 331348.
Ritter, J. R. (1991). The Long-Run Performance of Initial Public Oerings. Journal of
Finance 46.1, pp. 327.
Rosen, R. J. (2006). Merger Momentum and Investor Sentiment: The Stock Market Re-
action to Merger. Journal of Business 79.2, pp. 9871017.
Schlingemann, F. P. (2004). Financing Decisions and Bidder Gains. Journal of Corporate
Finance 10.5, pp. 683701.

25

Page 25 of 27
Schwert, G. (1996). Markup Pricing in Mergers and Acquisitions. Journal of Financial
Economics 41.2, pp. 153192.
Servaes, H. (1991). Tobins Q and the Gains from Takeovers. Journal of Finance 46.1,
pp. 409419.
Spiess, D. K. and J. Aeck-Graves (1995). Underperformance in Long-Run Stock Returns
Following Seasoned Equity Oerings. Journal of Financial Economics 38.3, pp. 243
267.
Travlos, N. G. (1987). Corporate Takeover Bids, Methods of Payment, and Bidding Firms
Stock Returns. Journal of Finance 42.4, pp. 943963.
Uysal, V. B. (2011). Deviation from the Target Capital Structure and Acquisition Choices.

t
Journal of Financial Economics 102.3, pp. 602620.

ip
Vermaelen, T. and M. Xu (2014). Acquisition Finance and Market Timing. Journal of
Corporate Finance 25, pp. 7391.

cr
Vladimirov, V. (2015). Financing Bidders in Takeover Contests. Journal of Financial
Economics 117.3, pp. 534557.
Walker, M. M. (2000). Corporate Takeovers, Strategic Objectives, and Acquiring-Firm

us
Shareholder Wealth. Financial Management 29.1, pp. 5366.

an
M
d
p te
ce
Ac

26

Page 26 of 27
- The paper directly investigates the financing source of takeovers

- Acquirers and targets characteristics influence a takeovers financing source

- Announcement returns for stock-financed takeovers underperform

- In the long run, all but stock-financed takeovers perform as expected

t
ip
cr
us
an
M
d
p te
ce
Ac

Page 27 of 27