Vous êtes sur la page 1sur 37

M&A and the Financial

Crisis
The influence of the 2008 Financial Crisis on deal volume, deal value
and average deal value in the United States

Lisa Tieleman
2603783

Thesis MSc Business Administration: Financial Management


School of Business and Economics
VU University Amsterdam
Supervisor: Prof. Dr. T.A. Gonzalez
29th of June, 2018
Abstract
This study examines whether M&A activity, more specifically deal volume, deal value and
average deal value, differs between the pre-crisis, crisis and post-crisis period. Regressions are
performed and controlled for firm-specific variables, time fixed effect, firm fixed effects and
robustness tests. The results showed that only the lower deal volume after the crisis in
comparison with before the crisis remained robust. Furthermore, the results indicate that before
the crisis, many deals were made with a great (average) deal value. When the crisis started, less
(average) deals were made, and the deal value was lower. After the crisis, deal volume still
declined but the (average) deal value became higher again. The results even indicate that after
the crisis, firms made bigger deals on average than before the crisis. It seems like after the
crisis, firms are more cautious about engaging in large investments such as M&A but since
cheap borrowing is available again, firms are willing to pay large amounts if they are convinced
of the benefits of the M&A.

Keywords: financial crisis; mergers and acquisitions; deal volume; deal value; average deal
value; probit model; OLS model

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 2
1. Introduction
The economic environment of the past decade is characterized by innovations and increasing
globalization. These developments cause changes like different customer demand such as more
digitalized products or increased foreign competition. Firms can adapt to these changes by
restructuring their operations to serve the demand or making investments abroad. A common
type of restructuring or investment is referred to as ‘mergers and acquisitions’ (hereafter
M&As). M&As represent substantial capital reallocations and have an estimated aggregate
volume of $1.34 trillion per year (Bonaime, Gulen, and Ion, 2017). Furthermore, M&As play
a vital role in the efficient allocation of capital. The deal usually creates synergies and enhances
shareholder value (Beeton, Ecko, and Thorburn, 2008; Sheen, 2014). Innovations and
increasing globalization partially caused the financial crisis that started in 2008 (Crotty, 2009;
Swagel, 2009). The impact of the financial crisis was tremendous and is seen as the most severe
financial crisis since the Great Depression in 1929 (Bordo and Meissner, 2015). It seems likely
that firms made different decisions during the 2008 financial crisis, including M&A deal
decisions.

M&A activity varies substantially across time and this variation has often been researched in
the past. However, most papers research the effects of the financial crisis compared with the
post-crisis period or the pre-crisis, also referred to as the sixth merger wave (2003-2007),
compared with the post-crisis period. Sánchez, Seeber and Golberg (2011) stated that M&A
activity was low during the financial crisis due to the financial turmoil and recovered slowly
after the financial crisis due to uncertainty about the economic outlook and tax changes. Reddy,
Nangia and Agrawal (2014) found that post-crisis results were lower than pre-crisis in number
of deals, deal value and average deal value for M&As. Narendar and Reddy (2015) replicate
the study from Reddy et al. (2014) and find similar results. Differences between these three
periods have not been examined yet in a single study with one dataset. This thesis contributes
to the discussion by examining M&A activity during the sixth merger wave, the financial crisis
and after the financial crisis and thus will examine the pre-crisis, crisis and post-crisis period.

In this study the effect of the financial crisis on M&As in the United States is examined. More
specifically, different characteristics of M&As such as deal volume, deal value and average
deal value are being researched. Several regressions are performed with M&A activity,
dummies for the different time periods and firm-specific control variables such as return on
assets (ROA), sales growth, firm size, leverage, liquidity and market-to-book equity ratio. The
regressions are performed again with firm and time fixed effects and a robustness check.
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 3
The main research question is formulated as followed:

“Does the value and frequency of M&As in the United States differ between the pre-crisis,
crisis and post-crisis period?”

The results showed that the deal volume was greater before the financial crisis than after the
financial crisis. Not as expected, the deal volume declined even more after the financial crisis.
This indicates that firms are being cautious with making large investments such as M&As. The
deal value and average deal value did not significantly differ during the different periods. The
deal value was highest before the financial crisis, lowest during the financial crisis and after
the financial crisis remained in between. The average deal value after the financial crisis was
bigger than before the financial crisis which was not as expected. This indicates that although
less deals were made, lending was cheap again and made it possible for firms to pay large sums.

This paper is organized as follows. Section 2 covers the literature review and the hypotheses.
Section 3 and 4 describes the methodology and data used to answer the hypotheses. Section 5
presents the results and makes conclusions about the hypotheses. Section 6 states a main
conclusion and discusses limitations and suggestions for further research.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 4
2. Literature review
Since the first merger wave occurred many academic papers have been written about M&A.
The first merger wave happened at the end of the 19th century. Afterwards, five other merger
waves occurred at the end of the 1920s, 1960s, 1980s, 1990s and the mid-2000s. This chapter
captures existing literature about M&As, more specifically deal value and deal volume and the
impact of the financial crisis on M&A deals. The following subjects are discussed: causes of
M&As, M&A activity during the pre-crisis period, M&A activity during the 2008 financial
crisis and M&A activity during the post-crisis period.

2.1 Causes of M&As


There are certain causes described in the literature that explain the reasoning behind M&As.
Trautwein (2013) distinguishes several theories that show the general motives to merge. The
most evidence has been found for the Valuation theory, the Empire-building theory and the
Process theory. The Valuation theory states that acquiring firms have better information about
the target firm than the market and can make better estimations. This is how acquirers can
benefit from undervalued target firms (Holderness and Sheehan, 1985). According to the
Empire-building theory, mergers are driven by managers who maximize their own utility
instead of shareholders’ value (Trautwein, 2013). Jensen’s (1986) free cash flow hypothesis
explains that firms, that have large free cash flow but few investment opportunities with a
positive net present value, will rather engage in unprofitable investments than returning cash
to shareholders. The Process theory claims that the decision to merge is driven by the strategic
decision process of a firm. This strategic decision process can be influenced by the limited
information processing ability of individuals. Roll’s managerial hubris hypothesis (1986) can
be put under Process Theory as well. This managerial hubris hypothesis states that managers
are overconfident when it comes to their own capabilities and overvalue synergies which causes
them to engage in M&A’s that will not be profitable. They overpay for the target firm at the
expense of the shareholder. Another reason for managers to merge or acquire is the need to
diversify their firm. Managers tend to do this when they are personally not well diversified
(Morck, Shleifer and Vishny, 1990). Managers will participate in diversifying that is not related
and will reason that a larger firm might be better protected against changes in the market and a
more diversified firm might be better protected against future challenges. Other theories which
are less plausible according to Trautwein (2013) are the Efficiency theory and the Monopoly
theory. The Efficiency theory explains mergers as being planned and executed to achieve

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 5
synergies. Brown and Sarma (2007) state that the created synergies will lead to a new combined
entity with a larger value than the two firms will have separately from each other. The
Monopoly theory views mergers as a way to achieve market power with a horizontal or vertical
M&A. A horizontal merger or acquisition occurs between firms that operate in the same
industry to achieve market share and synergies. A vertical merger or acquisition occurs when
a company merges with, or acquirers, one of its suppliers or customers.

These theories explain the internal motives for a firm to merge or acquire. These internal
motives can be influenced by external factors. These external factors can help trigger multiple
M&As which can result in what is called a merger wave.

2.2 M&A activity during the pre-crisis period


There are different theories described in the literature that can explain the reasoning behind
merger waves. A first reason could be external influences such as financial innovations,
deregulation and increased foreign competition (Mitchell and Mulherin, 1996). These changes
in the economic environment can cause severe economic shocks which can for example result
in a firm that cannot handle their increasing demand anymore. As a response to the change in
the economic environment, the firm has to restructure. It merges with or acquirers a firm with
similar capacities to extend their production capabilities. Lambrecht (2004) showed that
mergers arise in product markets that are rising. Mitchell and Mulherin (1996) researched the
US market from 1982 to 1989 and concluded that M&A is the most efficient way to cope with
economic shocks which is defined as a factor, expected or unexpected, that changes the
industry structure. Hartford (2005) studied whether M&A activity occurs due to market timing
or shocks and observed that for a merger wave to be present, growing capital- and stock markets
need to be observable and credit should expense rapidly. The study concluded that M&A waves
occur, in general, because of economic, technological or regulatory shocks but only when
sufficient access to capital is present. Hartford (2005) state that the occurrence of a merger
wave thus depends on the existence of a significant shock and the presence of liquidity.

Hartford (2005) did not find evidence for the market timing as a motive behind merger waves.
However, other studies have found evidence that market timing increases M&A activity in
times of economic expansion. According to Rhodes-Kroppf and Viswanathan (2004) a merger
wave can be caused by an overvalued market. Myers and Majluf (1984) state that during times
of financial boom, firms use their ‘overvalued’ stock to invest in other companies. Schleifer
and Vishny (1991) state that overvaluation across firms differ significantly during economic
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 6
expansion. Managers with less overvalued stock will purchase the more overvalued stock.
Martynova and Renneboog (2008) examined the pattern of M&A activity during different
merger waves and come to the conclusion that M&A waves are due to market timing from
managers but find only evidence for M&As that were paid for in stock.

Table 1 presents an overview of merger waves in the past, the causes of the merger waves and
the reasons why they ended. As can be seen, the causes and reasons to end M&A waves can be
placed under economical, technological or regulatory shocks.

Table 1: History of Merger Waves


Wave Period Cause of M&A wave End of M&A wave
Technological innovation;
first wave 1893-1904 WWI; stock market crash
economic expansion
Better antitrust laws; economic
second wave 1919-1929 The Great Depression
recovery
Economic recovery after WWII; Market crash because of oil
third wave 1955-1970
laws on anti-competitive M&A crisis
fourth wave 1974-1989 Deregulation; economic recovery Stock market crash
Deregulation; economic 9/11 attacks; burst of dot com
fifth wave 1993-2000
expansion bubble
Subprime mortgage crisis in the
sixth wave 2003-2008 Globalization; private equity
US

The sixth merger wave started in 2003 after the internet bubble. The economy was expanding
again, and financing rates were low which played a significant role in the M&A decision
(Alexandridis, Mavrovitis and Travlos, 2012; Hardford, 2005). Companies are looking for
ways to expand more than what they can achieve by internal or organic growth and turn to
M&As (Gaughan, 2009). As can be seen from Table 1, the sixth merger wave started due to
globalization and private equity (Metrick and Yasuda, 2011). Companies are looking to expand
beyond borders to create strong national or global champions and private equity grew due to
available debt financing and growing pools of investment capital (Lipton, 2006).

Martynova and Renneboog (2008) find that merger waves are interrupted by a large decrease
in stock markets and times of an economic crisis. This is confirmed by Gaughan (2009) as well.
The sixth merger wave indeed ended in 2008 when the US subprime mortgage crisis unfolded.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 7
2.3 M&A activity during the 2008 Financial Crisis
The 2008 financial crisis started with the collapse of the United States mortgage market (also
called subprime crisis) (Allen and Giovannetti, 2011; Boorman, 2009; Kamin and DeMarco,
2012) when the housing values dropped with 24% (Crotty, 2009). The financial crisis
developed further and the full extent of the exposure of financial institutions became apparent
and brought down leading investment banks such as Lehman Brothers. According to Gaughan
(2009) the bankruptcy of Lehman Brothers can be seen as the starting point of the 2008
financial crisis. The financial crisis is probably the largest crash of the world’s financial system
since the Great Depression in 1929 (Bordo and Meissner, 2015; Kowalski and Shachmurove,
2011), because of the globalized integration of capital markets. The financial crisis had a major
impact in all parts of the world due to this international character. Other factors that have
contributed to the financial crisis are the rise of securitization, growth of private capital and
increase in financial derivatives (Swagel, 2009). According to Crotty (2009) the combination
of deregulation and globalization of financial markets, in combination with financial
innovation, like securitization and derivatives, helped create conditions for this financial crisis.
The financial crisis caused M&A activity to decline as well (Gaughan, 2009; Ang and Mauck,
2011; Reddy et al, 2014).

Regarding the deal volume of M&As during the financial crisis, Gaughan (2009) states that the
deal volume of M&As declined. Reddy et al. (2014) studied cross border M&As and find the
same trend for 2008 and 2009. Banks stopped participating in the majority of the deals due to
the changing economic climate. Credit facilities became increasingly difficult to access and
caused higher costs of financing M&As. Sales and profits during a financial crisis often decline
which makes it unattractive to engage in a merger or acquisition. Some buyers waited with a
merger or acquisition until the capital market recovered and credit would be accessible again
at a decent rate. Other buyers purchased smaller targets or minority stakes as an alternative to
acquiring all of the target’s stock. In addition, Ang and Mauck (2011) studied the effects of fire
sales in the M&A market during a financial crisis. Results show that acquiring firms do not
benefit from acquiring distressed target firms. This is due to the misconception regarding the
benefits of fire sales discount which in reality does not hold. Ang and Mauck (2011) concluded
that less M&As occurred during the financial crisis.

Regarding the deal value of M&As during the financial crisis, Reddy et al. (2014) find a
decrease in deal value for 2008 and 2009 and stated that 2009 was largely affected by the global
financial crisis. Hotchkiss (1998) argued that firms in financial trouble are often purchased by
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 8
firms from the same industry that have had some sort of relationship with the distressed firm
in the past. The acquirer is aware of the true value of the target firm in bankruptcy and can
negotiate a more preferred deal. In contrast to Ang and Mauck (2011), Hotchkiss (1998) argues
that the deal value is at substantial discount compared to similar deals where the target firm is
not in bankruptcy. Acquirers have more leverage to negotiate a more desirable deal value
during a financial crisis, since target firms are more often in distress. Target firms are usually
smaller and thus more prone to have financial problems during a crisis and the acquirers have
a stronger position. Thus, it can be expected that the M&A deal value will be depressed during
the financial crisis.

2.4 M&A activity during the post-crisis period


According to Sánchez, Seeber and Golberg (2011), M&A activity slowly started to rise again
in the United States after 2010. Before the financial crisis, M&A was focused on the world’s
largest consumer markets which include the United States. M&A changed after the financial
crisis. Cross-border acquisitions are gaining popularity (Grave, Vardiabasis and Yavas, 2012;
Sánchez et al., 2011). Firms are looking for target firms in new geographies. According to
Gaughan (2014) debt financing has remained quite inexpensive, but this cheap lending was not
used to merge or acquire other firms. Companies choose to return cash to shareholders through
dividends and share repurchases and issue debt to refinance their prior costlier debt issues. In
the last four merger waves, including the sixth wave of 2003-2007, the economy was
expanding, and companies used M&As to bull market. Since stock prices were rising, potential
M&As were not being questioned. If the deal was put to question, the CEO argued that the
growth in equity value was due to their management. This has led to many M&A failings in
past merger waves. After the 2008 financial crisis CEOs reacted differently. Firms are more
cautious in terms of risks and expectations about the recovery of the economy and their own
firm. This led them to be extra careful when it comes to large investments such as M&As. The
uncertainty has led to a wider gap between what an acquirer is willing to offer and what a target
is willing to accept (Gaughan, 2014).

Reddy et al. (2014) find that post-crisis deal volume and deal value are comparatively lower
than pre-crisis, but the findings were not significant. Narendar and Reddy (2015) find the same
results. Furthermore, their post-crisis period sample consists of the years 2008, 2009 and 2010.
While 2008 and 2009 are the years in which the financial crisis unfolded, it can be questioned
whether these years should be considered as post-crisis years. Gaughan (2014) finds that the
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 9
deal volume rose significantly again in 2011 and 2012, but that the number of deals is nowhere
near its sixth-wave volume. Deals are still being formalized but just of smaller size and the
M&A market remains uncertain.

This study aims to answer the research question - Does the value and frequency of M&As in
the United States differ between the pre-crisis, financial crisis and post-crisis period? To
answer this research question, hypotheses are formulated which are based on the literature
review. They are as follows:

Hypothesis 1a: The M&A deal volume is less frequent during the financial crisis compared
with the pre-crisis period

Hypothesis 1b: The M&A deal volume is less frequent during the financial crisis compared
with the post-crisis period

Hypothesis 1c: The M&A deal volume is more frequent during the pre-crisis compared with
the post-crisis period

Hypothesis 2a: The M&A deal value is lower during the financial crisis compared with the pre-
crisis period
Hypothesis 2b: The M&A deal value is lower during the financial crisis compared with the
post-crisis period

Hypothesis 2c: The M&A deal value is greater during the pre-crisis compared with the post-
crisis period

Hypothesis 3a: The average M&A deal value is lower during the financial crisis compared with
the pre-crisis period

Hypothesis 3b: The average M&A deal value is lower during the financial crisis compared with
the post-crisis period

Hypothesis 3c: The average M&A deal value is greater during the pre-crisis compared with
the post-crisis period

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 10
3. Methodology
This chapter will explain the models and variables that are used to perform the empirical
analysis and help answer the hypotheses. This chapter will describe the regression
specifications and examine the dependent variables, independent variables and control
variables. It starts with a description of the model for the deal volume, continues with the model
for deal value and average deal value and ends with a description of the firm fixed and time
fixed effects. The process of gathering the data for the variables will be discussed in the next
chapter.

3.1 Deal volume


The empirical analysis starts with estimating the effects on the dependent variable deal volume.
The deal volume is measured as a dummy variable referred to as acquisition likelihood and
measures the likelihood of a firm being an acquirer in a given fiscal year. According to Keller
(2005) appropriate techniques to use with a dependent dummy variable is a probit or logistic
regression. This study follows the methodology of Bonaime et al. (2017) and will apply the
probit regression. The base regression is as follows:

AcquisitionDummyit = Φ (β0 + β1CrisisDummyit + β2Post-CrisisDummyit) (1)

The dependent variable in Equation (1) is the acquisition likelihood which equals 1 if firm i
made an acquisition during year t and 0 if otherwise and is referred to as AcquisitionDummyit.
The right-hand side contains the cumulative distribution function (Φ), the intercept (β0) and the
time period dummies CrisisDummyi and Post-CrisisDummyi. The Crisis dummy is an indicator
variable that equals 1 if t is a financial crisis year and equals 0 if otherwise. The Post-Crisis
dummy is an indicator variable that equals 1 if t is a post-crisis year and equals 0 if otherwise.
Pre-Crisis dummy, which equals 1 if t is a pre-crisis year and equals 0 if otherwise, is the
reference category and thus will not be included in the model. The pre-crisis period is defined
as the period from the 1st of January, 2005 until the 31th of December, 2007. The financial
crisis period is defined as the period from the 1st of January, 2008 until the 31st of December,
2010. The post-crisis period is defined as the period from the 1st of January, 2011 until the 31st
of December, 2013.

Equation (1) will be used to estimate an initial effect of the financial crisis compared with the
pre-crisis (Hypothesis 1a) and to estimate a first effect of the pre-crisis compared with the post-
crisis (Hypothesis 1c). A F test will be performed to test whether the financial crisis effect
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 11
differs from the post-crisis effect and the results will be used to draw initial conclusions on
Hypothesis 1b.

When control variables are added to the base regression model, the specifications become as
follows:

AcquisitionDummyit = Φ (β0 + β1CrisisDummyit + β2Post-CrisisDummyit + β3logSizesit +


β4ROAit + β5SalesGrowthit-1 + β6BookLeverageit + β7Liquidityit +
β8Market-To-Book-Equityit) (2)

The dependent variable, the cumulative distribution function (Φ), the intercept (β0) and the time
period dummies of Equation (2) are similar with Equation (1). Control variables are added to
see if firm-specific differences can help explain the dependent variable: log Size is the
logarithm of total book assets; ROA is earnings before extraordinary items plus interest expense
plus income tax divided by total book assets; Sales Growth is the sales in yeart minus sales in
yeart-1 divided by sales in yeart-1; Book Leverage is long-term debt plus debt in current liabilities
divided by total book assets; Liquidity is the cash to total assets ratio and is calculated by the
sum of cash and short-term investments divided by total book assets; and, Market-to-Book-
Equity is market value of equity divided by book value of equity.

Equation (2) will be used to estimate if the effect of the financial crisis compared with the pre-
crisis (Hypothesis 1a) and the effect of the pre-crisis compared with the post-crisis (Hypothesis
1c) still holds after controlling for firm-specific variables. A F test will be performed to test
whether the financial crisis effect differs from the post-crisis effect and the results will be used
to draw conclusions on Hypothesis 1b.

3.2 Deal value


The empirical analysis continues with estimating the effects on the dependent variable deal
value. The log of the deal value is taken because it improves the linearity of the model and
corrects for skewness in the data (Knopf, Nam and Thornton, 2002). According to Keller
(2005) the technique most appropriate with a dependent continuous variable and independent
variables that are either dummy or quantitative indicators is the Ordinary Least Square (OLS)
multiple regression analysis and is followed in this study. The base regression is as follows:

logDealValueit = α0 + β1CrisisDummyit + β2Post-CrisisDummyit + εit (3)

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 12
The dependent variable in Equation (3) is the log of the deal value that was paid for firm i
during year t. The right-hand side contains the intercept (α0), the time period dummies
CrisisDummyit and Post-CrisisDummyit and the error term (εit).

Equation (3) will be used to estimate an initial effect of the financial crisis compared with the
pre-crisis period (Hypothesis 2a) and the effect of the pre-crisis compared with the post-crisis
period (Hypothesis 2c) on the deal value. A F test will be performed to test whether the financial
crisis effect differs from the post-crisis effect and the results will be used to draw an initial
conclusion on Hypothesis 2b.

When control variables are added to the base regression model, the specifications become as
follows:

logDealValueit = α0 + β1CrisisDummyit + β2Post-CrisisDummyit + β3logSizeit + β4ROAit +


β5SalesGrowthit + β6BookLeverageit + β7Liquidityit +
β8Market-To-Book-Equityit + εit (4)

The dependent variable, the intercept (α0), the error term (εit) and the time period dummies of
Equation (4) are similar with Equation (3). Control variables are added to control for firm-
specific differences: log Size, ROA, Sales Growth, Book Leverage, Liquidity and Market-to-
Book-Equity.

Equation (4) will be used to estimate if the effect of the financial crisis compared with the pre-
crisis (Hypothesis 2a) and the effect of the pre-crisis compared with the post-crisis (Hypothesis
2c) on the deal value still holds after controlling for firm-specific variables. A F test will be
performed to test whether the financial crisis effect differs from the post-crisis effect and the
results will be used to draw conclusions on Hypothesis 2b.

3.3 Average Deal Value


The empirical analysis ends with estimating the effects on the dependent variable average deal
value. The logarithm of the average deal value is taken as well because of similar reasons as
with the deal value: it improves the linearity of the model and corrects for skewness in the data
(Knopf, Nam and Thornton, 2002). In line with the regression method for the deal value, the
regression analysis applied for the average deal value is the Ordinary Least Square (OLS)
multiple regression analysis (Keller, 2005) since the dependent variable is continuous and the
independent indicators either dummy or quantitative. The base regression is as follows:

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 13
logAverageDealValueit = α0 + β1CrisisDummyit + β2Post-CrisisDummyit + εit (5)

The dependent variable in Equation (5) is the logarithm of the average deal value that was paid
for firm i during year t. The right-hand side contains the intercept (α0), the time period dummies
CrisisDummyit and Post-CrisisDummyit and the error term (εit).

Equation (5) will be used to estimate an initial effect of the financial crisis compared with the
pre-crisis period (Hypothesis 3a) and the effect of the pre-crisis compared with the post-crisis
(Hypothesis 3c) on the average deal value. A F test will be performed to test whether the
financial crisis effect differs from the post-crisis effect and the results will be used to draw first
conclusions on Hypothesis 3b.

When control variables are added to the base regression model, the specifications become as
follows:

logAverageDealValueit = α0 + β1CrisisDummyit + β2Post-CrisisDummyit + β3logSizeit + β4ROAit


+ β5SalesGrowthit + β6BookLeverageit + β7Liquidityit +
β8Market-To-Book-Equityit + εit (6)

The dependent variable, the intercept (α0), the error term (εit) and the time period dummies of
Equation (6) are similar with Equation (5). Control variables are added to control for firm-
specific differences: log Size, ROA, Sales Growth, Book Leverage, Liquidity and Market-to-
Book-Equity.

Equation (6) will be used to estimate if the effect of the financial crisis compared with the pre-
crisis (Hypothesis 3a) and the effect of the pre-crisis compared with the post-crisis (Hypothesis
3c) on the average deal value still holds after controlling for firm-specific variables. A F test
will be performed to test whether the financial crisis effect differs from the post-crisis effect
and the results will be used to draw conclusions on Hypothesis 3b.

3.4 Firm fixed effects


To eliminate confounding factors, tests will be performed to improve the results. Firm fixed
effects help eliminate differences between firms which are unobserved and can affect the
results (Wooldridge, 2010). The sample of this study includes firms with different characters
from various industries and including firm fixed effects will help to control for systematic
deviations in risk and performance. Therefore, the OLS regressions will be tested for firm fixed
effects. According to Greene (2004) the nonlinear (i.e. probit) firm fixed effects model has two
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 14
shortcomings. A practical shortcoming, the difficulty of estimating a nonlinear model with
hundreds of dummy variable coefficients. A methodological shortcoming, the incidental
parameters problem which raises questions about the statistical properties of the estimator.
When the ratio of observations to number of parameters increases, parameter estimates will
converge to their true value because the standard errors decline. When fixed effects are added,
the number of parameters grow with the observations, thus standard errors increase. Therefore,
the probit model will not include firm fixed effects.

3.5 Time fixed effects


The aim of this study is to examine whether the financial crisis has an influence on M&A. To
make sure the results are only due to the financial crisis, other time-related factors have to be
excluded such as macroeconomic conditions. Therefore, time fixed effects will be included,
because it will help eliminate time-invariant confounding effects (Mummolo and Peterson,
2018).

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 15
4. Data
This chapter will explain how and from where the information data was collected. This study
uses two sources to gather the data, namely the Security Data Company (SDC) Platinum
database and the Compustat database. SDC Platinum is a database that collects data about
M&As in the United States. Compustat is a database that collects data about firm’s financial
statements.

4.1 Mergers and acquisition data


The sample consists of mergers and acquisitions in the United States between 2005 and 2013
from the SDC database. The dataset contains data regarding M&A announcements where the
companies are listed on a United States stock exchange and where the value of the transaction
is at least $1 million. The sample is limited to public companies because of the availability of
data. The acquiring firm must control less than 50% of the shares of the target firm before the
announcement date and must obtain 100% of the shares of the target firm after the transaction.
The SDC sample that fulfilled all the requirements contained of 1,858 deals. This sample was
matched with firm-specific information from the Compustat database where 954 of the deals
had data available. This dataset was cleaned by removing firms with data that was not complete
for every year or had illogical numbers such as similar deal values numerous times. When a
firm acquirers two or more companies in the same year, the total deal value of the deals will be
taken. Furthermore, following other M&A related studies (Beltratti and Paladino, 2013; Huang,
Jiang, Lie and Yang, 2014) to avoid the effect of outliers, the observations were winsorized at
the 1st and 99th percent, meaning that the highest and lowest 1% of the deals was removed from
the sample.

The final dataset consists of 491 deals with an average deal value of $2.7 billion. The lowest
deal in the dataset has a value transaction of $1 million and the largest deal value is $54.9
billion. The median deal value is $832 million, and the standard deviation is $5.8 billion. This
implies that the distribution of the sample has high variance and is skewed. To reduce the
skewness, the logarithm of the deal value is taken. The natural logarithm of deal value has the
following summary statistics: a mean of 6.568, a median of 6.723 and a standard deviation of
1.888. The natural logarithm of the average deal value is taken as well. The log of the average
deal value has the following summary statistics: a mean of 7.899, a median of 7.720 and a
standard deviation of 0.308.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 16
Table 2 presents the summary statistics of the total deal value and average deal value per year.
The total deal value varies from $80 billion to $289 billion where 2005 has the greatest total
deal value and 2012 has the lowest. The average deal value per year varies from $1.8 billion to
$4.8 billion where 2011 and 2005 have the greatest average deal value and 2012 the lowest.

Table 2: Total Deal value and Average Deal Value Summary Statistics
This table presents the summary statistics of the Average Deal Value and the Total Deal Value per year.
Total Deal Value, Average Deal Value and Std. Deal value are reported in millions of dollars $. Since
the data is skewed and has high variance, the table presents the logarithms of the Total Deal Value,
Average Deal Value and the Std. Deal Value as well.
log log Std.
# of Total Deal Average Deal Std. Deal log Total
Year Average Deal
deals Value Value Value Deal Value
Deal Value Value
2005 69 289,803.329 4,200.048 8,642.859 6.842 8.343 2.005
2006 68 208,733.203 3,069.606 5,347.636 6.962 8.029 1.610
2007 63 141,909.498 2,252.532 2,784.978 6.814 7.720 1.661
2008 51 114,232.541 2,109.305 6.838.881 5.647 7.654 2.193
2009 53 141,896.666 2,677.182 7,227.862 6.141 7.893 2.129
2010 60 133,163.037 2,174.467 3,762.765 6.622 7.685 1.707
2011 40 194,270.280 4,856.757 7.096.809 7.015 8.488 2.206
2012 44 81,193.868 1,845.315 2.181.939 6.664 7.520 1.637
2013 43 88,982.989 2,069.372 2,961.779 6.774 7.635 1.541

4.2 Control variables


The equation as described in chapter 3 includes six control variables. Following Bonaime et al.
(2017) this study uses the following: logarithm of the total assets, return on assets (ROA), sales
growth, book leverage, cash holdings and market-to-book equity. All firm-level control
variables were computed with the following items which were gathered through the Compustat
database:

● Total assets (Compustat annual item AT)


● Income before extraordinary items (Compustat annual item IB)
● Interest expense (Compustat item XINT)
● Income taxes (Compustat item TXT)
● Sales (Compustat annual item SALE)
● Long term debt (Compustat annual item DLTT)
● Debt in current liabilities (Compustat item DLC)
● Cash and short-term investments (Compustat item CHE)

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 17
● Share price (Compustat annual item PRCC_F)
● Common shares outstanding (Compustat item CSHO)
● Common equity (Compustat item CEQ)
● Carrying value of preferred stock (Compustat item PSTK)

The database Compustat is used to get the firm-specific variable data. The financial statements
of the previous year have to be taken when the merger or acquisition has not taken place yet.
Thus, the sample starts with all firms that have available data for the period 2004 until 2012
since the research period is from 2005 until 2013. The sample consisted of 18,766 firms and
112,548 observations. The dataset will be cleaned by deleting firms that did not have all the
data for every year in the sample and duplicates. After deleting the data 2,962 firms will remain
where 372 firms have done at least one acquisition during the sample period and 2,590 firms
do not do any acquisition during the sample period. Every firm has nine years of observations
which means that the corresponding number of observations must be the amount of firms times
these nine years. Table 3 presents the summary statistics of the control variables.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 18
Table 3: Summary Statistics for Firm-Specific Control Variables
This table presents summary statistics on firm-specific control variables which are derived from the Compustat
database. The data contains firm-year observations from 2004 until 2012. Panel A reports the acquirers which
are identified as firms who have made an acquisition at least once during the sample period. Panel B report the
non-acquirers who are identified as firms who have not made any acquisition during the sample period. Together,
they represent the full sample of 26,658 observations of 2,962 firms (Panel C). ROA represents return on assets
in percentages. Size is measured as the logarithm of total assets. Sales growth is represented in percentages as
the growth of sales measured against prior year growth of sales. Book leverage is measured as long-term debt
plus current liabilities, divided by total assets. Liquidity represents the cash to total assets ratio and is calculated
as cash divided by total assets. Market-to-book equity is the market to book equity ratio and is calcuated as
the market value of equity divided by the book value of equity.

Mean Std. Dev. P10 Median P90

Panel A: Acquirers
ROA 0.071 0.139 -0.011 0.078 0.179
Size 8.513 2.037 5.984 8.538 10.910
Sales growth 0.103 0.246 -0.095 0.066 0.349
Book leverage 0.265 0.217 0.019 0.225 0.538
Liquidity 0.129 0.148 0.010 0.077 0.324
Market-to-book equity 2.671 4.029 0.839 2.042 5.294
Number of observations 3,720
Panel B: Non-Acquirers
ROA 0.010 0.335 -0.174 0.063 0.181
Size 6.403 2.556 2.940 6.551 9.631
Sales growth 0.085 0.288 -0.191 0.062 0.372
Book leverage 0.277 0.373 0.000 0.211 0.578
Liquidity 0.161 0.193 0.008 0.086 0.436
Market-to-book equity 2.308 4.629 0.510 1.723 5.145
Number of observations 25,900
Panel C: All firms
ROA 0.018 0.317 -0.146 0.065 0.181
Size 6.401 2.546 2.944 6.520 9.605
Sales growth 0.089 0.284 -0.182 0.065 0.372
Book leverage 0.276 0.357 0.000 0.213 0.571
Liquidity 0.156 0.187 0.008 0.085 0.419
Market-to-book equity 2.354 4.553 0.545 1.766 5.168
Number of observations 29,620

Panel A of Table 3 presents the summary statistics for the acquiring firm, Panel B presents the
summary statistics for the non-acquiring firms and Panel C presents the summary statistics for
all firms. The variables seem to differ substantially from each other. For example, the size of
acquirers is 8.51 and the size of non-acquirers is 6.40 which indicates that acquirers tend to be
larger firms than non-acquirers. Furthermore, the market-to-book equity of acquirers and non-
acquirers are 2.67 and 2.31 which indicate that the equity of the firms are overvalued in general:
acquirers even more so than non-acquirers. To test whether these differences are statistically
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 19
significant, t-tests for all control variables are performed. The results of the t-tests can be found
in Table 4.

Table 4: t-test of Acquirers and Non-Acquirers


This table presents the t-test to test whether the acquirer sample (Panel A) significantly differs from the non-
acquirer sample (Panel B). The last two columns P2.5 and P97.5 represent the 95% confidence interval. *,
** and *** indicate statistical signifance at the 10%, 5% and 1% level, respestively.
Mean Std. Err.
t-test P2.5 P97.5
difference difference
Panel A: T-test between Acquirers and Non-Acquirers
ROA 0.061 0.006 10.471*** 0.050 0.073
Size 2.111 0.046 46.022*** 2.021 2.201
Sales growth 0.018 0.005 3.520*** 0.008 0.091
Book leverage -0.011 0.007 -1.781* -0.025 0.001
Liquidity -0.032 0.003 -9.317*** -0.039 -0.025
Market-to-book equity 0.363 0.084 4.327*** 0.199 0.528
Degrees of freedom 26,658

The results of the t-tests confirm the expectations that the acquirers financial statement
variables statistically differ from the non-acquirers. Acquirers have a larger ROA which
indicates that they are more profitable in general. Acquirers are larger in size and have a higher
sales growth. Moreover, acquirers have a higher market-to-book equity ratio which indicates
that their stock is even more overvalued than the stock of non-acquirers. Acquirers have lower
book leverage which indicates they have less debt. Acquirers are less liquid which indicates
that they have less cash holdings in comparison with non-acquirers.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 20
5. Results
In this chapter, the results of the empirical analysis to estimate the effect of the financial crisis
on M&A activity measured as deal volume, deal value and average deal value are presented.
The effects on each dependent variable are discussed for the different time periods; pre-crisis,
financial crisis and post-crisis. The hypotheses will be rejected or not.

The effects on the deal volume is measured with a probit regression where the deal volume is
measured as a dummy which equals one if a firm made an acquisition in a given year and zero
otherwise. This dummy is referred to as acquisition likelihood and the marginal effects at the
mean are calculated. This means that if all other variables are hold at their means, the marginal
coefficient represents the increase or decrease in the probability of acquisition likelihood. This
indicates that if the marginal coefficient is positive, the probability of acquisition likelihood is
greater and thus a higher deal volume can be expected and vice versa.

The effects of the deal value and average deal value are measured with an OLS regression
where the logarithm of deal value and average deal value are taken. Since these dependent
variables are logs, the exponentiation of the coefficients has to be calculated to provide
interpretation, because it is the inverse of the logarithm function and represents the proportional
difference (Manning, 1998). The results will provide the beta coefficients and the
exponentiation of the beta coefficients which will be used for interpretation. The exponentiated
beta coefficients of an independent dummy variable indicates that the dependent variable will
be the given coefficient times a hundred percent higher or lower as compared with the reference
group. The exponentiated beta coefficient of a continuous independent variable indicates that
a ten percent increase will increase or decrease the dependent variable with the given
coefficient times ten percent. The exponentiated beta coefficients of an independent log
variable indicates that a difference of one in the independent log variable results in the given
coefficient increase or decrease in the dependent log variable.

5.1 Effects on deal volume


In Figure 1 as presented below, the deal volume over the years is plotted. As can be seen from
the graph, the deal volume remains stable for several years (2005-2007) after which a large
drop occurs. The deal volume remains stable again after which another drop occurs. The
decrease in deal volume after the financial crisis period (especially from 2010 to 2011) is not
in line with the previous literature as described in this paper.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 21
The results of the empirical analysis will reveal if this observed decline is statistically
significant.

Number of deals
80
70
60
In numbers

50
40
30
20
10
0
2005 2006 2007 2008 2009 2010 2011 2012 2013
Year

Number of deals

Figure 1: Number of deals


This figure show the movement of the amount of acquisitions conducted by US
public firms from January 2005 until December 2013.

Table 5 presents the results of the marginal effects at the mean of the financial crisis on
acquisition likelihood. As can be seen, the effects are relatively small. Model (I) shows that the
probability of an acquisition taking place during the financial crisis is 0.4% lower when
compared with the pre-crisis period which is significant at the 5% significance level. The
probability of an acquisition taking place during the post-crisis period is -0.7% lower when
compared with the pre-crisis period which is significant at the 10% significance level. The F
test indicates that the marginal effects of the financial crisis and post-crisis significantly differ
from one another. This implicates that deals will most likely take place before the financial
crisis, then during financial crisis and that deals will least likely take place after the financial
crisis. The number of deals is decreasing from beginning to end of the sample period. This is
in line with what was observed from Figure 1.

After controlling for different firm-specific variables (II), the marginal effects of both the
financial crisis (-0.3%) and the post-crisis (-0.6%) remain robust and are significant at the 5%
and 1% level, respectively. The F test indicates that these marginal effects statistically differ
from one another. This indicates that after controlling for firm-specific variables, deals will
still most likely take place before the financial crisis and least likely after the financial crisis.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 22
Size, liquidity and ROA have significant positive marginal effects of 0.5%, 1.0% and 0.8%,
respectively. This indicates that firms that are larger, which are more profitable and have more
cash holdings, are slightly more likely to acquire or merge with another firm. These kinds of
firms have a strong position and money thus can outbid other firms who are smaller in size,
have less money and lower profitability. Book leverage has a negative significant marginal
effect of -0.9%. This indicates that firms with debt are slightly less likely to acquire or merge
with another firm. A target may rather be merger with or acquired by a firm with less debt.
Market-to-book-equity has a significant positive marginal effect but is negligible.

After controlling for different firm-specific variables and time fixed effects in model (III), only
the marginal effect of the post-crisis (-0.5%) compared with the pre-crisis remains robust and
is significant at the 5% level. The F test indicates that the marginal effects of the financial crisis
and post-crisis do not statically differ from each other. This indicates that after controlling for
firm-specific variables and time fixed effects, deals will least likely take place after the
financial crisis and that there is no difference in number of deals between the pre-crisis and
financial crisis. All control variables remain similar except sales growth which becomes
significant (0.5%), indicating that firms with higher sales growth have a slightly higher
probability of engaging in M&A within time. The results indicate that after controlling for
different firm-specific variables and time fixed effects, the number of deals is once again
significantly higher before the financial crisis than after the financial crisis and that this effect
is not due to time factors.

A Likelihood Ratio test is performed to check for heteroscedasticity in the model and resulted
in the following: χ2 = 21.27, p = 0.007. This indicates that model (III) is biased by
heteroscedasticity. The regression is performed again with robust standard errors to control for
the bias. The results of model IIII remain similar and the F test that tests whether the financial
crisis effect differs from the post-crisis effect remains insignificant. This means that Hypothesis
1a and 1b can be rejected, the number of deals does not differ between the pre-crisis and
financial crisis, and between the financial crisis and post-crisis. This is not in line with Ang
and Mauck (2011). The post-crisis effect compared with the pre-crisis effect remains robust
through all different regressions. This indicates that the number of deals before the financial
crisis is significantly higher than after the financial crisis which means Hypothesis 1c cannot
be rejected. This is in line with Gaughan (2014). Firms seem to be cautious with making M&A
deals even when the crisis is over.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 23
Table 5: Results for the Acquisition Likelihood
This table presents results from the probit regressions of the acquisition likelihood on dummies for the CRISIS and
POST-CRISIS (I) that equal one if the acquisition is made during the crisis or after the crisis and equals zero if
otherwise, respectively. (II) controls for the firm-specific variables. SIZE is measured as the logarithm of total assets
and ROA represents return on assets in percentages. SALES GROWTH is represented in percentages as the growth
of sales measured against prior year growth of sales. BOOK LEVERAGE is measured as long-term debt plus current
liabilities, divided by total assets. LIQUIDITY represents the cash to total assets ratio and is calculated as cash divided
by total assets. MTB EQUITY is the market to book equity ratio and is calcuated as the market value of equity divided
by the book value of equity. (IIII) controls for time fixed effects, firm-specific variables and adds a robustness test to
control for heteroskedasticity. dydx represents the marginal effect at the mean and standard errors are reported in
paratheses. The bottom of the table presents the results from a F test to test whether the crisis and post-crisis effect
differ from one another. *, ** and *** indicate statistical significance at the 10%, 5% and 1% level, respectively.
Acquisition likelihood I II III IV
Variable dy/dx dy/dx dy/dx dy/dx
Compared with PRE-CRISIS
CRISIS -0.004** -0.003 -0.001 -0,001
(0.002) (0.001) (0.002) (-0.002)
POST-CRISIS -0.007*** -0.006*** -0.005** -0.005**
(0.002) (0.001) (0.002) (0.002)
SIZE _ 0.005*** 0.005*** 0.004***
(0.000) (0.000) (0.000)
ROA _ 0.008* 0.008** 0.008**
(0.004) (0.004) (0.003)
SALES GROWTH _ 0.003 0.005** 0.005**
(0.002) (0.002) (0.002)
BOOK LEVERAGE _ -0.009*** -0.009*** -0.009***
(0.003) (0.003) (0.003)
LIQUIDITY _ 0.010** 0.010** 0.010**
(0.005) (0.004) (0.003)
MTB EQUITY _ 0.000** 0.000*** 0.000***
(0.000) (0.000) (0.000)

Firm fixed effects NO _ NO _ NO _ NO _

Time fixed effects NO _ NO _ YES _ YES _


Robustness check NO _ NO _ NO _ YES _
Observations 29,610 29,610 29,610 29,610

test F p-value F p-value F p-value F p-value


Crisis – Post-Crisis = 0 2.99 0.0839* 5.88 0.015** 2.63 0.105 2.83 0.102

5.2 Effects on deal value


In Figure 2 as presented below, the total deal value over the years is plotted. As can be seen
from the graph, the deal value is the highest in 2005 and decreases the following years. The
deal value increased in 2009 again which is not as expected and decreased in 2010. After 2010,
the deal value starts increasing significantly again. The second high is in 2011 after the financial
crisis, but then the total value decreases again which is not as expected.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 24
Deal value
350
In billions of dollars $ 300
250
200
150
100
50
0
2005 2006 2007 2008 2009 2010 2011 2012 2013
Year

Deal value

Figure 2: Deal value


This figure shows the movement of the deal value of acquisitions conducted by US
public firms, from January 2005 until December 2013 in billions of dollars.

Table 6 presents the results with the beta coefficients and the exponentiated beta coefficients,
of the effects of the financial crisis on the deal value. According to the empirical analysis, the
deal value during the financial crisis is 51.0% lower than during the pre-crisis, which is
statistically significant at the 1% level. This indicates that firms make deals with a lower value
during the financial crisis than before the financial crisis. The effects of the financial crisis
compared with after the financial crisis statically differs at the 1% significance level according
to the F test. This indicates that when the financial crisis was over, firms made deals with higher
value again. The deal value does not statistically differ between the pre-crisis period and the
post-crisis period which indicates that the deal values before and after the financial crisis are
similar.

After controlling for the firm-specific variables, the financial crisis effect declines (to 45.1%)
but remains robust and statistically significant at 1% level in comparison with the pre-crisis
period. This indicates that after controlling for firm-specific variables, the deal value is still
lower during the financial crisis than before the financial crisis. The deal value before and after
the financial crisis still do not statistically differ from one another. The F test remains
significant at the 1% level, which indicates that the deal value during the financial crisis is
significantly lower than after the financial crisis after controlling for firm-specific variables.
Size, ROA and sales growth have a significant positive effect on the deal value. If size, which
is a log variable, increases with a difference of one, the deal value will increase with 51.7%. If
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 25
ROA and sales growth for example increase with ten percent the deal value will increase with
2.74% and 7.43%, respectively. This indicates that larger firms, who are more profitable and
have higher sales growth, tend to make deals with a higher value. These firms have more money
and thus the ability to acquire target firms with greater value.

When controlling for firm-specific variables and fixed effects, all effects become insignificant.
The deal value does not statistically differ before, during and after the financial crisis. The size
and sales growth positive effects on the deal value remain robust which indicate that time
factors and differences in firms do not have an effect.

A Breusch-Pagan test is performed to test for heteroscedasticity and resulted in the following:
χ2 = 4.10, p = 0.0428. The regression will be performed once more since the test is significant
at the 5% level. After adding a robustness check and controlling for firm-specific variables and
fixed effects, all variables are insignificant, including size and sales growth. The F test remains
insignificant. This indicates that after controlling for firm-specific variables, firm fixed effects
and time fixed effects and adding a robustness check, the deal value does not significantly
differ between the periods. This is not in line with other studies (Gaughan, 2014; Hotchkiss,
1998). Furthermore, none of the control variables have an effect. This indicates that the deal
value does not differ before, during and after the financial crisis which means Hypothesis 2a,
2b and 2c can be rejected.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 26
Table 6: Results for the Deal Value
This table presents the results with the beta coefficients (β) and the exponentiation of the beta coefficients (exp β) from
the OLS regressions of the logarithm of the deal value on dummies for the CRISIS and POST-CRISIS (I) that equal
one if the acquisition is made during the financial crisis or after the financial crisis and equals zero if otherwise, respectively.
(II) controls for the firm-specific variables. SIZE is measured as the logarithm of total assets and ROA represents return
on assets in percentages. SALES GROWTH is represented in percentages as the growth of sales measured against
prior year growth of sales. BOOK LEVERAGE is measured as long-term debt plus current liabilities, divided by total
assets. LIQUIDITY represents the cash to total assets ratio and is calculated as cash divided by total assets. MTB
EQUITY is the market to book equity ratio and is calcuated as the market value of equity divided by the book value
of equity. (III) controls for firm-specific variables and time and firm fixed effects. (IIII) controls for firm-specific variables,
time and firm fixed effects and adds a robustness test to control for heteroskedasticity. The bottom of the table presents
the results from a F test to test whether the CRISIS and POST-CRISIS effect differ from one another. Standard errors
are reported in paratheses and *, ** and *** indicate statistical significance at the 10%, 5% and 1% level, respectively.
Deal Value I II III IV
Variable β exp β β exp β β exp β β exp β
Compared with PRE-CRISIS
CRISIS -0.713*** -0.510 -0.600*** -0.451 -0.390 -0.323 -0.401 -0.330
(0.189) (0.189) (0.290) (0.399)
POST-CRISIS 0.034 0.035 -0.064 -0.062 -0.127 -0.119 -0.128 -0.120
(0.200) (0.200) (0.313) (0.525)
SIZE _ 0.417*** 0.517 0.410*** 0.507 0.407 0.502
(0.043) (0.043) (0.312)
ROA _ 0.242*** 0.274 0.247 0.280 0.247 0.280
(0.075) (0.475) (0.311)
SALES GROWTH _ 0.556* 0.743 0.593** 0.809 0.592 0.808
(0.330) (0.289) (0.582)
BOOK LEVERAGE _ 0.734 1.083 0.746 1.109 0.744 1.104
(0.436) (0.471) (1.861)
LIQUIDITY _ 0.462 0.587 0.484 0.623 0.484 0.623
(0.609) (0.611) (2.033)
MTB EQUITY _ -0.014 -0.014 -0.008 -0.008 -0.008 -0.008
(0.020) (0.020) (0.031)

Firm fixed effects NO _ NO _ YES _ YES _


Time fixed effects NO _ NO _ YES _ YES _
Robustness check NO _ NO _ NO _ YES _
Adjusted R2 0.029 0.271 0.083 0.083
Observations 491 491 491 491

test F p-value F p-value F p-value F p-value


Crisis – Post-Crisis = 0 12.09 0.001*** 7.70 0.006*** 0.64 0.425 1.07 0.344

5.3 Effects on average deal value


In Figure 3 as presented below, the average deal value over the years is plotted. As can be seen
from the graph, the average deal value is more than $4 billion in 2005 and decreases the
following years. The average deal value increased in 2009 which is not as expected and
decreased in 2010. After 2010, the deal value starts increasing significantly again. The highest
average deal value is in 2011 after the financial crisis, but then the total value decreases again
___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 27
which is not as expected. The movements of the average deal value are similar to the
movements of the deal value.

Average deal value


6

5
In billions of dollars $

0
2005 2006 2007 2008 2009 2010 2011 2012 2013
Year

Average deal value

Figure 3: Average deal value


This figure shows the movement of the average deal value of acquisitions conducted
by US public firms, from January 2005 until December 2013 in billions of dollars.

Table 7 presents the result of the effect of the financial crisis on the average deal value, with
beta coefficients and exponentiated beta coefficients. According to the empirical analysis, the
average deal value during the financial crisis is 25.6% lower and the post-crisis average value
is 15.7% in comparison with the pre-crisis average deal value which is significant at the 1%
level. The F test is significant as well which indicates that the financial crisis and post-crisis
effects are statistically different from one another. The results indicate that the average deal
value is highest before the financial crisis, lowest during the financial crisis and after the
financial crisis is in between. This is in line with the literature.

When controlling for the firm-specific variables, the effect of the financial crisis in comparison
with the pre-crisis remains robust. The average deal value is lower during the financial crisis
than before the financial crisis. The magnitude of the effect has declined to from -25.6% to
-13.2%. The F test remains significant which indicates that the effect of the financial crisis in
comparison with the post-crisis remains robust. The average deal value is still lower during the
financial crisis than after the financial crisis. The average deal value before and after the
financial crisis do not significantly differ anymore. None of the control variables have
significant effect on the average deal value.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 28
When controlling for firm-specific variables and fixed effects, the financial crisis effect in
comparison with the pre-crisis becomes insignificant and the F test is no longer significant.
Remarkably, the post-crisis effect changes into a positive effect which indicates that the
average deal value is 18.4% greater after the financial crisis than before the financial crisis.
This is in contradiction with what was expected (Gaughan, 2014). Size is the only significant
control variable after controlling for firm-specific variables and fixed effects. If size, which is
a log variable, increases with a difference of one, the average deal value will decrease with
22.3%. This is noteworthy, because it seems like larger firms make deals with less value than
smaller firms on average. It may indicate that because the firms are larger in size, they can
negotiate a better deal, or they acquire more small firms.

A Breusch-Pagan test is performed to test for heteroscedasticity and resulted in the following:
χ2 = 70.46, p = 0.000. The regression will be performed once more since the test is significant
at the 1% level. After controlling for firm-specific variables and fixed effects and adding a
robustness check for heteroscedasticity, all effects are insignificant and thus the average deal
value after the financial crisis is no longer significantly greater than before the financial crisis.
This means Hypothesis 3a, 3b and 3c can be rejected. The size effect remains similar and the
market-to-book-equity ratio becomes significant but small. It indicates that overvalued firms
tend to have a slightly lower average deal value.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 29
Table 7: Results for the Average Deal Value with exponentiated beta coefficients
This table presents the results with the beta coefficients (β) and the exponentiation of the beta coefficients (exp β) from
the OLS regressions of the logarithm of the average deal value on dummies for the CRISIS and POST-CRISIS (I) that
equal one if the acquisition is made during the financial crisis or after the financial crisis and equals zero if otherwise,
respectively. (II) controls for the firm-specific variables. SIZE is measured as the logarithm of total assets and ROA
represents return on assets in percentages. SALES GROWTH is represented in percentages as the growth of sales
measured against prior year growth of sales. BOOK LEVERAGE is measured as long-term debt plus current liabilities,
divided by total assets. LIQUIDITY represents the cash to total assets ratio and is calculated as cash divided by total
assets. MTB EQUITY is the market to book equity ratio and is calcuated as the market value of equity divided by the
book value of equity. (III) controls for firm-specific variables and time and firm fixed effects. (IIII) controls for firm-
specific variables, time and firm fixed effects and adds a robustness test to control for heteroskedasticity. The bottom of
the table presents the results from a F test to test whether the CRISIS and POST-CRISIS effect differ from one another.
Standard errors are reported in paratheses and *, ** and *** indicate statistical significance at the 10%, 5% and 1%
level, respectively.
Average Deal Value I II IV IV
Variable β exp β β exp β β exp β β exp β
Compared with Pre-Crisis
CRISIS -0.296*** -0.256 -0.141*** -0.132 -0.074 -0.071 -0.074 -0.071
(0.030) (0.032) (0.068) (0.051)
POST-CRISIS -0.171*** -0.157 -0.008 -0.008 0.169* 0.184 0.169 0.184
(0.032) (0.033) (0.091) (0.120)
SIZE _ -0.000 -0.000 -0.252*** -0.223 -0.252*** -0.223
(0.007) (0.090) (0.090)
ROA _ -0.055 -0.054 -0.322 -0.275 -0.322 -0.275
(0.120) (0.612) (0.407)
SALES GROWTH _ -0.011 -0.011 -0.001 -0.001 -0.001 -0.001
(0.054) (0.136) (0.123)
BOOK LEVERAGE _ -0.119 -0.112 -0.474 -0.377 -0.474 -0.377
(0.074) (0.365) (0.308)
LIQUIDITY _ -0.014 -0.014 0.106 0.112 0.106 0.112
(0.097) (0.430) (0.276)
MTB EQUITY _ -0.002 -0.002 -0.009 -0.009 -0.009** -0.009
(0.003) (0.008) (0.004)

Firm fixed effects NO _ NO _ YES _ YES _


Time fixed effects NO _ NO _ YES _ YES _
Robustness check NO _ NO _ NO _ YES _
Adjusted R2 0.171 0.171 0.002 0.002
Observations 491 491 491 491

test F p-value F p-value F p-value F p-value


Crisis – Post-Crisis = 0 13.83 0.000*** 17.36 0.000*** 0.64 0.425 1.27 0.386

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 30
6. Conclusion
This study found that M&A activity does differ between the pre-crisis, financial crisis and post-
crisis period, but not all effects remain robust. Deal volume is greater before the financial crisis
than after the financial crisis and the after-crisis deal volume is even lower than during the
financial crisis. Deal value and average deal value were initially greatest before the financial
crisis, lowest during the financial crisis and after the financial crisis was in between. However,
the effects did not remain robust. The post-crisis average deal value seemed even higher than
the pre-crisis average deal value, although not significant.

The results showed that the number of deals is not lower during the financial crisis in
comparison with before and after the financial crisis. The results showed as well that the
number of deals before the financial crisis is higher than after the financial crisis. As can be
seen from Figure 1, the deal volume shows a consistent pattern with a drop between 2007-2008
and 2010-2011. This can indicate that since the financial crisis started, and the sixth merger
wave ended, firms are more careful with M&A. Since cash holdings, size, ROA and sales
growth had positive effects, it can help explain that there were still firms engaging in M&As
that were profitable, had cash balance, sales that grew and were big in size. Not as expected,
the number of deals declined even more after the financial crisis. This can implicate that after
the financial crisis, firms are more cautious with engaging in large investments such as M&A.
Gaughan (2014) argued that although lending was cheap again after the crisis, this was not used
to engage in M&A but rather to return cash to shareholders and issue debt to refinance prior
costlier debt. The financial crisis was different than the ones before because of the international
character and was the greatest one since the Great Depression in 1929. These facts can help
explain why deal volume did not become greater again after the financial crisis.

The results showed that the M&A deal value does not differ before, during and after the
financial crisis. The M&A deal value was initially lower during the financial crisis than before
or after, but the effects disappeared after controlling for fixed effects. This indicates that the
crisis effect is significant on average, but not within firms and over time. Bigger firms that have
higher sales growth initially had greater M&A deal value, even within firms and over time.
These effects disappeared after a robustness check. This indicates that an initial effect was
present, but the model might suffer from omitted variable bias. As can be seen from Figure 2,
the deal value declined, but took a jump in 2011 where after it declined again.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 31
The results showed that the average M&A deal value does not differ before, during and after
the financial crisis. Initially, the average deal value was lowest during the financial crisis, then
after the financial crisis and highest before the financial crisis, as expected. However, the
results did not remain robust after controlling for fixed effects and robustness test. This
indicates that the average deal value was on average greatest before the financial crisis, lowest
during the financial crisis and after the financial crisis was in between, but not within firms and
over time and the model might suffer from omitted variable bias. Remarkably, the results
indicated that the average deal value was greater after the financial crisis than before within
firms and over time. This is in contradiction with expectations. As can be seen from Figure 3,
the average deal value has a similar pattern as the deal value. However, the deal value in 2011
is lower than before the financial crisis, but the average deal value is higher than before the
financial crisis. This indicates that after the financial crisis, firms make less deals, but of greater
value than before the financial crisis.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 32
7. Discussion
This thesis studied the effects of M&A activity before, during and after the financial crisis. The
results of this study have some implications for M&A consultants and firms. However, this
study has limitations that might have influenced the results and need to be considered.
Suggestions for further research are made that can cope with these limitations.

First, this study investigated 491 deals over a nine-year period which indicates that about 55
deals per year were researched. This study found that deal volume after the financial crisis was
even lower than during the financial crisis. Further research could use a larger dataset and find
out if this result and others hold or might be due to the small dataset in this study.

Since some effects turned insignificant after checking for heteroscedasticity, this might indicate
omitted variable bias. Furthermore, the models explained as much as 27.1% of the variation.
Further research could add more variables to see if they help explain more of the variation and
reduce the omitted variable bias.

Third, further research could research different sample periods. For example, the financial
crisis period contained data from January 1th of 2008 on, but the US subprime crisis started
before this date. This can indicate that during the pre-crisis sample there were signs of a
financial crisis present and could have influenced the results. The financial crisis sample
contained 2010 but it can be questioned whether this year may belong to the post-crisis period.
The increasing deal volume in 2010 confirms this questioning.

This study does not control for firm fixed effects in the probit regression, because of the
incidental parameter problem. Advanced methods as proposed by Fernández-Val and Weidner
(2016) who use analytical and jackknife bias corrections to correct for this problem and get
unbiased estimators could be applied to give more robust results.

Even though it seems likely that firms are more cautious with making deals after the financial
as it was the greatest since the Great Depression, it seems noteworthy that deal volume after
the crisis was even lower than during the crisis. A reasoning for this could be the increasing
interest in cross-border M&A. To get a full picture, further research could extent the sample to
worldwide M&As.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 33
8. References
Alexandridis, G., Mavrovitis, C. F., and Travlos, N. G. (2012). How have M&As changed?
Evidence from the sixth merger wave. The European Journal of Finance, 18(8), 663-
688.

Allen, F., and Giovannetti, G. (2011). The effects of the financial crisis on Sub-Saharan
Africa. Review of Development Finance, 1(1), 1-27.

Beitel, P., and Schiereck, D. (2001). Value creation at the ongoing consolidation of the
European Banking Markets.

Beltratti, A., and Paladino, G. (2013). Is M&A different during a crisis? Evidence from the
European banking sector. Journal of Banking & Finance, 37(12), 5394-5405.

Betton, S., E. Eckbo and K. Thorburn. (2008). Corporate takeovers. Handbook of Corporate
Finance: Empirical Corporate Finance 2: 291–430.

Bonaime, A. A., Gulen, H., and Ion, M. (2017). Does Policy Uncertainty Affect Mergers and
Acquisitions?

Boorman, J. (2009). The impact of the financial crisis on emerging market economies: The
transmission Mechanism, policy response and lessons. In Global Meeting of the
emerging Markets Forum.

Bordo, M. D., and Meissner, C. M. (2015). Financial Globalisation, Financial Development


and Financial Crises in the Golden Age, 1870-1914. In conference “Financial
Systems and Economic Growth: Conference in Honor of Richard Sylla” Stern School
of Business, New York University March (pp. 27-28).

Brown, R., and Sarma, N. (2007). CEO overconfidence, CEO dominance and corporate
acquisitions. Journal of Economics and business, 59(5), 358-379.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 34
Capron, L. (1999). The long-term performance of horizontal acquisitions. Strategic
management journal, 987-1018.

Crotty, J. (2009). Structural causes of the global financial crisis: a critical assessment of the
‘new financial architecture’. Cambridge journal of economics, 33(4), 563-580.

Fernández-Val, I., and Weidner, M. (2016). Individual and time effects in nonlinear panel
models with large N, T. Journal of Econometrics, 192(1), 291-312.

Gall, E. A. (1991). Strategies for merger success. Journal of Business Strategy, 12(2), 26-29.

Gaughan, P. A. (2009). M&As in troubling times. Journal of Corporate Accounting &


Finance, 20(2), 45-50.

Gaughan, P. A. (2014). US M&A Market Lags Corporate Performance. Journal of Corporate


Accounting & Finance, 25(2), 11-18.

Grave, K., Vardiabasis, D., and Yavas, B. (2012). The global financial crisis and M&A.
International Journal of Business and Management, 7(11), 56.

Greene, W. (2004). The behaviour of the maximum likelihood estimator of limited dependent
variable models in the presence of fixed effects. The Econometrics Journal, 7(1), 98-
119.

Harford, J., (2005). What drives merger waves? Journal of Financial Economics, vol. 77, pp.
529-560.

Holderness, C. G., and Sheehan, D. P. (1985). Raiders or saviors? The evidence on six
controversial investors. Journal of Financial Economics, 14(4), 555.

Hotchkiss, E. S., and Mooradian, R. M. (1998). Acquisitions as a means of restructuring firms


in Chapter 11. Journal of Financial Intermediation, 7(3), 240-262.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 35
Huang, Q., Jiang, F., Lie, E., and Yang, K. (2014). The role of investment banker directors in
M&A. Journal of Financial Economics, 112(2), 269-286.

Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. The
American economic review, 76(2), 323-329.

Kamin, S. B., and DeMarco, L. P. (2012). How did a domestic housing slump turn into a
global financial crisis?. Journal of International Money and Finance, 31(1), 10-41.

Keller, G. (2005). Statistics for management and economics (Seventh edition ed.). Duxbury.

Knopf, J. D., Nam, J., and Thornton, J. H. (2002). The volatility and price sensitivities of
managerial stock option portfolios and corporate hedging. The Journal of
Finance, 57(2), 801-813.

Kowalski, T., and Shachmurove, Y. (2011). The financial crisis: What is there to learn?
Global Finance Journal, 22(3), 238-247.

Lambrecht, B. M. (2004). The timing and terms of mergers motivated by economies of


scale. Journal of Financial Economics, 72(1), 41-62.

Lipton, M. (2006). Merger Waves in the 19th, 20th and 21st Centuries. The Davies Lecture,
Osgoode Hall Law School, York University, 14, 21.

Manning, W. G. (1998). The logged dependent variable, heteroscedasticity, and the


retransformation problem. Journal of health economics, 17(3), 283-295.

Martynova, M., and Renneboog, L. (2008). A century of corporate takeovers: What have we
learned and where do we stand?. Journal of Banking & Finance, 32(10), 2148-2177.

Mitchell, M. L., and Mulherin, J. H. (1996). The impact of industry shocks on takeover and
restructuring activity. Journal of financial economics, 41(2), 193-229.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 36
Morck, R., Shleifer, A., and Vishny, R. W. (1990). Do Managerial Objectives Drive Bad
Acquisitions?, 45(1), 31–48.

Mummolo, J., and Peterson, E. (2018). Improving the Interpretation of Fixed Effects
Regression Results. Political Science Research and Methods, 1-7.

Myers, S. C., and Majluf, N. S. (1984). Corporate financing and investment decisions when
firms have information that investors do not have. Journal of financial
economics, 13(2), 187-221.

Ossadnik, W. (1996). AHP-based synergy allocation to the partners in a merger. European


Journal of Operational Research, 88(1), 42-49.

Roll, R. (1986). The hubris hypothesis of corporate takeovers. Journal of business, 197-216.

Rhodes-Kroppf, M., and Viswanathan, S., (2004). Market Valuation and Merger Waves. The
Journal of Finance, vol. 59, pp. 2685-2718.

Sánchez, C., Seeber, C. D., and Goldberg, S. R. (2011). M&A update. Journal of Corporate
Accounting & Finance, 22(2), 9-13.

Sheen, A. (2014). The real product market impact of mergers. The Journal of Finance 69:
2651–2688.

Shleifer, A., and Vishny, R. W. (1991). Takeovers in the '60s and the '80s: Evidence and
Implications. Strategic management journal, 12(S2), 51-59.

Wooldridge, Jeffrey M. (2010). Econometric Analysis of Cross Section and Panel Data. MIT
Press.

___________________________________________________________________________
Master Thesis – L.M. Tieleman – The effects of the financial crisis on mergers and acquisitions 37

Vous aimerez peut-être aussi