To buy or not to buy, this is here the question.
The influence of mergers and acquisitions on the innovativeness of
pharmaceutical companies in
the global economic up- and downturns of 1997 - 2008
Master Thesis
By
Ivo Tokarski s2013347
Supervisors
Dr Killian McCarthy
Dr Rene van der Eijk
University of Groningen
Faculty of Economics and Business
Master of Science in Business Administration
Strategy & Innovation
27
thJuly 2012
Abstract
This research gives insights on the issue to what extent mergers and acquisitions (M&As)
of global pharmaceutical companies influence their innovativeness in economic up- and
downturns. In total 1615 pharmaceutical companies were investigated through an
empirical analysis. Evidence for three main findings has been found. First, pharmaceutical
companies that acquire in slight economic downturns have higher positive returns.
Secondly, the number of patents that have been traded does not influence the returns of
the company - therefore the innovative performance does not increase measurably.
Thirdly, the location of US based target firms has a positive effect on the number of patents
that have been traded.
Preface
The pharmaceutical industry always has interested me personally, since it is an industry
that has a lot of power to it. As I saw the chance to write my master thesis on this topic I
did not hesitate and grabbed this opportunity. I was keen to do an empirical research that
is more about number crunching than interpreting the statements of other researchers. A
big challenge on my way, as a MScBA Strategy & Innovation student, was to combine
empirical hard data calculation with vague and not really exact variables describing
innovativeness. Nevertheless, I am happy with the outcome of the work because in a way
together with my supervisor we have created something new from a vast puzzle of
numbers as the starting point. In this respect, I would like to thank my supervisor dr
Killian McCarthy, who was more than just support on my way to graduation. I feel like I
have learned a lot from him and I am very thankful for his dedication. Also, I would like to
thank Ilona Faryna, who helped me immensely to stay on track and motivated me
throughout the process.
Index
Abstract ... 2
Preface ... 3
1. Introduction ... 5
2. Theoretical reflection ... 6
2.1 M&A influence ... 6
2.2 Innovativeness ... 7
2.2.1 Innovation and the moderator: Structural inertia ... 8
3. Methodology ... 9
3.1 Sample ... 9
3.2 Dependent variables ... 11
3.3 Independent variables ... 12
3.4 Control Variables ... 13
3.5 Descriptive statistics ... 13
3.6 Calculation Model ... 14
4. Results ... 15
4.1 Summary of findings – univariate regression... 15
4.2 Summary of findings - multivariate regression ... 16
4.3 Summary of findings - multinominal logistic regression ... 16
5. Discussion ... 18
5.1 Patents ... 18
5.2 Periods ... 19
5.3 Locations ... 22
5.4 Limitations ... 24
Summary ... 24
References ... 25
Internet sources... 27
Appendix ... 27
A – Correlation matrix ... 27
B - Univariate Regression ... 29
C - Multivariate Regression ... 47
1. Introduction
In the pharmaceutical industry companies are reliant upon scale and innovativeness to
develop products and stay profitable by covering their expenses for R&D. To stay
competitive and profitable the companies require therefore large cash reserves. This
needed affluence in cash reserves is restrained when a recessionary period strikes the
global economy. By investigating how those companies adapt their M&A behaviour to
these new economical circumstances and measuring what is affecting their innovativeness,
will provide managerial insights on which strategy could pay-off in the end.
Managers and CEO’s of pharmaceutical companies are dealing with a competitive
surrounding in which they have to guide their enterprises through the increasing rivalry
within their industry sector caused by new technologies, pioneering research and
innovation to secure the next blockbuster drug. An additional major concern is the threat
of appearing substitutes, which could replace the patented pharmacological products and
therefore cut the companies existing commercialization chain - leaving them behind on
huge R&D costs, which cannot be compensated for anymore. Furthermore, the economic
cycles create another challenge in raising funds and marketing the products to keep the
innovativeness up. Those companies and their leaders have to find new ways to secure
their innovativeness on the one hand, and block out their competitors on the other. An
established method that pharmaceutical firms seem to rely on is mergers and acquisitions
(M&As), in order to acquire valuable patents to increase their potential for innovativeness,
as well as a method to block out competition strategically.
Previous researchers like Pisano (1997) found out that based on transaction cost theory,
M&As compared to alliances, have lower transaction costs in cases that involve insecure
property rights and, transaction-specific production goods and when a transfer of complex
technology is needed. Basing on the study of Pisano (1997) the pharmaceutical industry
follows a logical pattern using M&As as a form of collaboration that have a low transition
cost.
Contrasting the idea of transaction cost theory of the pharmaceutical industry towards
industry non-specific studies of Kohers and Kohers (2000), shows that M&As in general
create slightly negative shareholder value, which gives an indication that even when
transaction costs are low – no value is created. Supporting this statement is the study of
Moeller et al. (2005), who studied the merger wave of 1998 – 20001 and found out that
more value was destroyed than created within this period.
This research will conduct therefore an empirical analysis of the global pharmaceutical
industry, in order to investigate how the innovativeness of these acquiring companies is
affected by mergers and acquisitions - in the context of economic up- and downturns.
The scope of a global perspective shall have the purpose to distinguish if the number of
patents as an innovativeness indicator has an influence on various other performance
indicators, which could give important managerial insights.
acquires a smaller company the influence on innovativeness is measureable (Bouncken,
2011; Giovannetti et al., 2011) and should have a negative effect on the small company
itself (Christensen, 2006). But what effect does it have on the acquiring company?
To set a researchable timeframe I chose the fifth and sixth global merger waves, which
took place in between 1997 – 2008 (see figure 1). This gives the chance to distinguish two
economic upturns and two economic downturns, creating a basis for comparison and
validity.
This classification creates an opportunity to establish how the M&As of pharmaceutical
companies affected the innovativeness of the acquiring companies within this closed time
period that had economic upturns as well as economic downturns.
2. Theoretical reflection
To successfully conduct this research, two main factors have to be put into theoretical
perspective when defining them, namely ‘Patents’ and ‘Periods’ - as this research needs to
establish measurement criteria for innovativeness (indicated through patents), and
economic up- and downturns given by two upturn and two downturns periods.
A big part of the scientific literature believes that measuring innovativeness in any
industry is difficult per se, because of the fuzzy logic behind it. When it comes to
measurements like R&D spending’s and similar innovativeness indicators it is rather
suitable to compare a small amount of companies that have no diverse demographics.
When taking a global approach to examine whether there is a general pattern for the
worldwide pharmaceutical industry, innovativeness has to be defined by a common
measure that is internationally independent. Therefore, the number of patents that have
been traded to the acquiring firm by M&A, shall distinguish the transition of
innovativeness (potential) from the target company, to the acquirer. Further, this research
is focused on examining whether this transition of patents from the target company to the
acquirer, has performance benefits for the acquiring company in economic up- or
downturns.
The timeframe is set within 1997 – 2008 because it gives additional insights of the
performance in two economic upturn periods and two economic downturn periods.
The balancing act of the acquirer in this setting is described as a need to innovate in order
to stay in business, which on one hand should focus on not restricting the innovation
process, but on the other hand should guide to preferred destinations (Pal, 2010). This
means that there should be incentives to foster innovativeness but in a preferable
direction for later commercialization. If it is not possible however, the here researched
M&As give the opportunity to acquire innovativeness through patents, which fit into the
company’s product pipeline.
2.1 M&A influence
The motives for M&As can be classified overall into adaptive or defensive rationales,
versus proactive or offensive rationales (Burns et al., 2005). For the industry context of
pharmaceutical companies those classifications do also apply because some M&As were
performed in order to fill gaps in a company’s product pipeline to maintain growth
(Grabowski et al., 2002), so were rather defensive in nature, whereas other M&As were
focused on the proactive rationale of increasing the scale, scope and R&D productivity
(Cockburn & Henderson, 2001).
Further implications for M&As in the pharmaceutical industry are related to a study of the
authors Danzon et al. (2007), who found out that firms with a relatively old portfolio of
marketed drugs exhibit a higher propensity to acquire another firm. It can be argued that
economic stress is influencing the M&A activity, which should well relate to the economic
up- and downturns.
In order to measure the influence of M&As this study will adopt the idea of the authors
Arza and Lopez (2011), who showed that companies linked to research organisations
invest more in innovative activities and are more prone to patenting – which gives a
theoretical hint of the linkage between the number of links and innovativeness. Therefore,
this research shall define the ‘M&A influence’ as the amount of linkages and how often they
have been created between the acquiring company and any given target, in other words –
the frequency of M&As that have been performed in a given period. This shall be one of the
dependent variables of the statistical analysis that should answer what influence the
frequency of M&As have on the company’s innovativeness.
Interestingly, there are different motives and frequencies of M&As when comparing global
pharmaceutical firms. The research of Demirel and Mazzucato (2010), has found evidence
that the location of the most innovative US based pharmaceutical firms is associated with
their location. Taking this into account and extending it to a global perspective the
question arises whether the different locations of the targets that have been acquired,
effects the innovativeness of the acquiring company differently. Also the question of
whether demographic areas had different strategies to cope with the fluctuations of the
global pharmaceutical market and their implications can be additional insights of this
study and broadening the focus of patents and periods by adding another variable.
Therefore following hypotheses shall be tested:
H1: The location of the target firm has an effect on the number of patents that have
been traded.
2.2 Innovativeness
Further, Lanjouw and Schankerman (2002) have researched the patent quality and
research productivity by measuring innovation with multiple indicators. One of them was
the quality of patents that allowed making a judgement on how productive a company was
in terms of R&D. For the research of pharmaceutical companies this measurement criteria
should be changed into the amount of applied patents for each given company. As for the
other central factors, this innovativeness indicator gives grounds to compare
pharmaceutical companies in a global context, without having to constrict the results due
to the complexity of the different geographical environments those companies encounter.
Interpreting the research of Lanjouw and Schankerman further, this patenting frequency
should be positively related to the stock market value of firms. The stock market value is
on the other hand related to the financial performance of the firm, which connects the
innovativeness to the above mentioned performance factor and should also in some way
relate to the M&As of the firm.
The pharmaceutical industry has its own implications regarding the measurement of
innovativeness, because it heavily relies on patenting - were the patent virtually equals the
product (Thomas, 2003). Within this industry the patented products could be easily and
cheaply reproduced by competitors, when this protection barrier would not exist. And
since the capital investment in creating a marketable drug is exceptionally high and has to
pass several R&D stages and FDA approval, patent protection is crucial to innovative
companies in order to retain a guaranteed period of market exclusivity, which should give
the opportunity to recoup the development costs (Bale & Harvey, 1997). For that reason
my research will adopt the number of applied patents that have been traded from the
target company to the acquirer as a variable to measure the potential increase of
innovative output of the acquiring company.
Following hypothesis shall therefore also be tested:
H2: The number of patents that have been transferred from the target to acquire have
a positive effect on innovative performance.
2.2.1 Innovation and the moderator: Structural inertia
This seems to be the major opposition force towards the transaction cost theory (Pisano,
1997), which seems to lose validity as structural inertia increases.
In global downturns many businesses experience an increased uncertainty within their
industry due to the turbulence in the financial market (Sahin et al., 2011). Even though the
pharmaceutical industry seems to be a special case, since there is no recessionary period
for illness and need of medication, they nevertheless will be affected by the reduced
financial flow caused by the recession. The capital-intensive business model of large
pharmaceutical firms is based on high R&D expenditures with only a few products to
recoup the investment that went into the development and marketing of the products. As
the large pharmaceutical companies were ‘lumbering’ ahead in their early growth phases
to attain scale economies and high turnovers, they were sufficient in covering their own
investments in product developments (even though financial leverage through debt
and/or outside investors was appreciated because of risk reduction and tax savings). As
those companies grew in times of prosperities, they were able to put some financial ‘meat’
on their structural bones, allowing them to survive economic downturns and global
recessionary periods for a certain time. Smaller biotech/pharma companies caught by a
global recession in their early stages of development do not have the means to attract
financial investors and loans in this economic climate of uncertainty, which varies strongly
over time – with uncertainty levels rising by 50 percent to 100 percent during recession
(Bloom, 2007). Companies with little cash reserves will find themselves soon in the
position of standing with their back against the wall, with the only chance of survival by
being acquired through a large(er) pharmaceutical company that has got the necessary the
cash reserves. The problematic situation for the smaller companies is that they will not
have any bargaining power towards the acquiring firm, since their economic survival is
dependent on the bigger firms. For the large pharmaceutical companies on the other hand
it seems like a global economic downturn is like a private invitation of the top performers
of the industry to go on a ‘cheap shopping tour’, buying out prospecting companies along
with their patents in order to increase their product portfolio. The question for the
acquiring companies should be whether they should integrate those M&As or leave them
separate by providing only financial support, since the study of McCarthy and Weitzel
(2009) has shown that compared to large firms, acquiring SMEs, are more flexible and
more able to avoid deals that turn sour. This new strategic implications could show that
bridging the financial needs of a start-up in order to let it grow by itself, could be more
profitable in the long run, than vertical integration.
Following hypothesis shall therefore be tested:
H3: Pharmaceutical companies that acquire in economic downturns have higher
positive returns.
3. Methodology
3.1 Sample
The initial raw data sample was extracted through the database SDC, which included all
mergers and acquisitions that have been published. The date of announcement for the
mergers and acquisitions was narrowed down to the timespan of 01/01/1997 to
31/12/2008. Acquiring companies were narrowed down to companies that possess the
primary SIC Code of 2822, 2834, 2835 and 2836. Further, the deal type code (1) was
included, as well as the deal status code (C) and the acquirer public status code (P). In total
n=1615 deals could be extracted.
expression. The data for the financial performance indicators covers 1137 of the 1615
M&A deals.
Additionally, the data on patents has been counted manually by researching the European
Patent Office database (EPO), and counting the patents that have been filed until the day of
the M&A. In total patent data for 970 companies could be extracted.
The performance indicators for this paper will be set into the context of financial
indicators that will allow to make a ‘hard’ data based analysis.
This research shall base on one of the most extensively studied areas, namely the financial
component of how economic goals of the company are fulfilled (Barney, 2002). It is rather
not beneficial to use the extended model of performance measurement because it gives to
complex results for the regression analysis in order to analyse the pharmaceutical
companies in global terms.
The financial perspective is defined by the tangible outcomes of the strategy using
traditional financial terms, like economic value added, revenue growth, costs, profit
margins, cash flow, net operating income, etc (Grigoroudis et al., 2012). In respect to that
in this study I will also define further dependent variables influencing innovativeness in
terms of cumulative abnormal returns of a given company. There are several reasons for
that. Firstly, the availability of data for different measurement indicators like e.g.
‘economic value added’ or ‘profit margins’ are hard to obtain, and mostly not publically
available. The return index of pharmaceutical companies on the other hand is available
because most of the companies are listed on the stock exchange and therefore need to
publish these numbers. Secondly, the cumulative abnormal returns indicate the ability to
capture value from the commercialization of products, which gives a variable that includes
several indicators of whether the product or innovation is commercially successful from
the beginning of the R&D to marketing. Third and lastly, it also gives common grounds to
compare companies on a global level because regional taxes are excluded and the focus
lays just on the commercial success of income generation. Therefore, cumulative abnormal
returns shall be the other main variables in defining the performance measurement
statistics, which goes along with the authors Dehning and Richardson (2002) who showed
in their research synthesis, on how to calculate returns on investment in a complicated
industry like IT, that a wide range of performance measurements can be used, including
event studies based on the shareholder return, on stock performance and sales growth
percentage ratios.
The research setting for this study is defined as the global pharmaceutical industry. In
total 1615 companies were included which acquired other companies within the
timeframe of 1997 – 2008. This timeframe was divided into four periods that are defined
as economic upturns (Period 1, Period 3) and economic downturns (Period 2, Period 3).
The division of into those four periods corresponds with the fifth and sixth global merger
wave of pharmaceutical companies (mergerstat.com, 2012).
3.2 Dependent variables
The following section presents the dependent variables that have been chosen to
distinguish the behaviour of M&As in terms of frequency, in terms of the change in
percentage of returns to the shareholders in long-term, the acquiring companies stock
price value change in long-term, as well as a short-term measurement of the cumulative
abnormal return. These dependent variables have the purpose of indicating if there are
performance indications in long-term, short-term or in relation to the activity in M&A
pursuance of the acquiring company.
Frequency of M&As (freq_ma)
The amount of mergers and acquisitions that have been performed within the specific
period, in other words the frequency of M&As, is set to be the dependent variable.
According to the authors Cassiman et al. (2005) M&As affect performance and
organizational output. Setting this dependent variable into contrast to the independent
and control variables, should give insights about how the activity of acquirers influence
the other factors. It should give valuable strategic insights of whether an active or rather
passive M&A behaviour has better implications on later performance and innovativeness.
The data could be obtained by counting every M&A the acquirer performed in a given
period, therefore covering n=1614 companies. This variable is additionally a good
indicator of how risk averse an acquirer is and if it has do with economic up- and
downturns, or locations when controlling for them.
Total Return Index – Abnormal return growth (tri_ar)
The abnormal return growth rates of the total return index are calculated on the basis of
CAR (-3, -1) Quartiles, as the estimation window, which has been set into contrast to the
CAR (0, +1) Quartiles as the event window. This adjusted growth rate has been
benchmarked to the growth rates of the total return index global pharmaceutical industry.
The t-test shows significance in comparison of the population means with an indicator of
0,26.
The abnormal return growth rates of the total return index, is a measurement to show to
which extent shareholder value has been created. This measurement shows if the M&A in
the end has created shareholder value in the long-term.
All data has been extracted from Datastream and accounts for n=1137 of the acquiring
companies.
Figure 2 – Abnormal return growth rates, Total Return Index
-400 -200 0 200 400 600 800 1000 Abn or ma l Reu tr n Per cent age 1997-2008
Price Index – Abnormal return growth (pi_ar)
The abnormal return growth rates of the price index of the acquiring company has been
calculated as the above-mentioned variable (tri_ar) with CAR (-3, -1) Quartiles as the
estimation window and CAR (0, 1) Quartiles as the event window. The adjusted abnormal
return has been benchmarked to the price index of the global pharmaceutical industry.
The t-test shows also significance at an indicator of 0,47.
This variable is a more pessimistic version of the (tri_ar) variable, and displays to which
extent the stock price of the company has been affected through the M&A, in respect to
growth percentages. This measurement shows whether the company itself has profited
from the M&A in the long-term. The coverage is also n=1137 of the in total n=1615
acquiring companies
Figure 3 – Abnormal return growth rates, Price Index
Cumulative Abnormal Return (car21_1)
An additional dependent variable is calculated by extracting the cumulative abnormal
return on a daily basis of -21 days before the M&A, as the estimation window,
benchmarking the average to the industry standard, and in the end contrasting this
performance to +1 day after the M&A. This variable has the reason to distinguish the
companies’ performance on daily basis by examining the return index and not on a
quarterly basis like the variables (tri_ar) and (pi_ar). Further, this variable does not
investigate the growth rates but the financial performance of the given company and
therefore is a suitable measure to distinguish the firms’ returns in the short-term. The
coverage of this variable is n=1136 of in total n=1615 acquiring companies.
3.3 Independent variables
Number of patents (num_pat)
The main independent variable for this research is the number of patents that has been
traded as the acquiring firm merged or acquired the target company. The number of
patents has been extracted by checking each individual target company in the European
Patent Office database, and counting the patents they have possessed until the target
company has been acquired. The patents that have been traded by M&A from the target
company to the acquirer are set to be an indicator of the potential for innovativeness that
has been transferred to the acquiring company.
3.4 Control Variables
In general, this study is controlling for periods, location and what influence the periods
with locations have on the innovativeness given by the number of patents.
Locations (row_tar), (us_tar), (eu_tar), (ni)
Each Location has been coded with a dummy variable in order to distinguish the areas in
which the target companies have been acquired. The variable (us_tar) are US based
targets, (eu_tar) distinguish the European based target and all other countries are covered
through the variable (row_tar), which stands for the rest of the world. Additionally the
variable (ni) is giving insight whether the deals are national or international which
provides information about the acquirer.
Periods (p1), (p2), (p3), (p4), (u), (d), (tp)
When controlling for periods the deals have been coded with a dummy variable for each
period in that the acquisition took place. Ranging from p1 to p4 the individual periods
cover the timeframe of all mergers and acquisition within 1997 – 2008.
(p1) is defined as the period from 1997-1999, (p2) is 2000-2002, (p3) is 2003-2006 and
(p4) is 2007-2008. Additionally periods are coded as economic up- and downturns, in
which (p1) and (p3) display upturns given by the variable (u), and the periods (p2) and
(p4) are described by the variable (d).
Further, turning points (tp) defined as peak performances in the abnormal return growth
rates are included into the model by an established event window of -1 Quartile and +1
Quartile from the given max/min of the (tri_ar) and (pi_ar).
3.5 Descriptive statistics
Variable
Obs Mean
Std. Dev.
Min Max
usp1
1615 1.092879
.2903531
1
2
usp2
1615 1.117647
.3222895
1
2
usp3
1615 1.182663
.3865093
1
2
usp4
1615 1.09969
.2996796
1
2
eup1
1615 1.040248
.1966004
1
2
eup2
1615 1.056966
.2318493
1
2
eup3
1615 1.086687
.2814634
1
2
eup4
1615 1.063158
.2433221
1
2
ltri_ar
1615 1.063777
.2444311
1
2
lpi_ar
1615 1.064396
.2455336
1
2
lcar21_1
1138 1.093146
.2907645
1
2
lfreq_ma
1615 1.091022
.287729
1
2
(For the correlation matrix please see Appendix A.)
3.6 Calculation Model
The basis calculation model of this study is determined by finding the influence of the
number patents that have been transferred from the target company to the acquirer by
merger or acquisition, on four different performance indicators. The performance
indicators are given by the dependent variables, namely, (car21_1), (tri_ar), (pi_ar) and
(freq_ma).
The calculation includes three calculation models on which each dependent variable is
tested individually for each control variable. The control variables are divided into control
variables for periods, as well as control variables for locations. The independent variable is
defined as the number of patents.
First an univariate regression analysis is being performed on which the individual
performance shall be indicated. The calculation model for the univariate analysis contains:
Dependent = Patents
Dependent = Periods
Dependent = Locations
Dependent = Patents x Periods
Dependent = Patents x Locations
Dependent = Patents x Periods x Locations
The second calculation is a multivariate analysis, in which the four dependent variables,
indicating the performance, are controlled by the same independent and control variables
as mentioned above (Dependent = Independent + Control).
The third calculation model is a multinominal logitistic regression, which adds the
likelihood ratios to the analysis. The performance indicators given by the four dependent
variables have been coded displaying positive (coded as 1) or negative (coded as 0)
relationships for the dependent variables (car21_1), (tri_ar), (pi_ar). For the dependent
variable (freq_ma) the weighted average has been set as a benchmark and if the frequency
of mergers and acquisitions for a given company scored above the average, coded as 1, and
if it underperformed the average it has been coded as 0. In the dataset those coded
variables are called (lcar21_1), (ltri_ar), (lpi_ar) and (lfreq_ma), respectively.
LDependent = Patents x Periods
LDependent = Patents x Locations
LDependent = Patents x Periods x Locations
4. Results
This section contains the summarized result tables of the regression analyses in sequence
of univariate regression, multivariate regression and multinominal logistic regression. For
each block the main results are presented. For the complete calculation output please see
Appendix B, C and D.
4.1 Summary of findings – univariate regression
The univariate regression shows that overall the frequency of M&As is lower in economic
downturns than in upturns, with negative implications for financial returns at turning
points. Further, it shows that acquiring US based targets in economic downturns has a
positive influence on the growth of the stock prize of the acquiring company, as well as for
the growth in shareholder value.
Model Variables Dependent Significance Observations Calculation
1 d freq_ma -0.223* 1614 Periods
2 eu_tar pi_ar, num_pat -52.61** 701 Patents x Locations 3 eu_tar tri_ar, num_pat -46.67* 701 Patents x Locations
4 eu_tar pi_ar -44.15** 1137 Locations
5 eup2 freq_ma, num_pat -0.692* 970 Patents x Periods x Locations 6 eup2 car21_1, num_pat -0.692* 970 Patents x Periods x Locations
7 p1 pi_ar -64.61** 1137 Periods
8 p1 pi_ar, num_pat -59.93* 701 Patents x Periods
9 p2 pi_ar, num_pat 47.36* 701 Patents x Periods
10 p2 pi_ar 34.72* 1137 Periods
11 p2 freq_ma, num_pat -0.632*** 970 Patents x Periods
12 p2 freq_ma -0.407** 1614 Periods
13 p3 freq_ma, num_pat 0.353** 970 Patents x Periods
14 p3 freq_ma 0.328** 1614 Periods
15 p4 pi_ar -40.30* 1137 Periods
16 rowp3 pi_ar, num_pat 58.49* 701 Patents x Periods x Locations
17 tp car21_1 29.99* 1136 Periods
18 tp tri_ar, pat -87.44*** 701 Patents x Periods
19 tp tri_ar -59.96*** 1137 Periods
20 u freq_ma 0.223* 1614 Periods
Those findings indicate that acquiring in moderate economic downturns, has the best
financial implications for the acquirer, especially when they are based in the US. Avoiding
turning points, which indicate peak performances in the market, could lower the risk of
fluctuations in the market. Further, acquiring companies in moderate economic downturns
could mean to get a cheaper deal when performing M&As (see figure 5 and 6).
4.2 Summary of findings - multivariate regression
The multivariate regression shows that when the frequency of mergers and acquisition
increases, also the number of patents that have been traded increases. Further it confirms
the pattern of the univariate regression that turning points in the market should be
avoided, because they account for highly negative performances, especially in generating
shareholder value.
Model Variables Dependent Significance Observations Calculation
1 d pi_ar -44.98** 827 Periods
2 eu_tar pi_ar, num_pat -67.86** 511 Patents x Locations 3 eup3 car21_1, num_pat -75.67* 511 Patent x Periods x Locations
4 num_pat freq_ma 0.00251* 511 Patents
5 num_pat freq_ma, num_pat 0.00252* 511 Patents x Locations 6 num_pat freq_ma, num_pat 0.00250* 511 Patent x Periods x Locations
7 p2 car21_1 53.26* 827 Periods
8 p2 freq_ma, num_pat -0.821** 511 Patents x Periods
9 p3 pi_ar 44.98** 827 Periods
10 p4 pi_ar -37.85* 827 Periods
11 row_tar car21_1, num_pat 50.66* 511 Patents x Locations 12 row_tar pi_ar, num_pat 58.01** 511 Patents x Locations 13 rowp3 car21_1, num_pat 71.18* 511 Patent x Periods x Locations 14 rowp3 pi_ar, num_pat 71.02** 511 Patent x Periods x Locations
15 tp tri_ar -77.94*** 827 Periods
16 tp pi_ar -39.45* 827 Periods
17 tp tri_ar, num_pat -100.0*** 511 Patents x Periods
18 u pi_ar 44.98** 827 Periods
19 usp2 freq_ma, num_pat -0.955* 511 Patent x Periods x Locations
Those findings indicate that a high frequency of M&As indeed increases the patent flow to
a company, which could increase their potential of generating innovations – but on the
other hand no measurable financial performance indicator can be associated to the
amount of patents a company acquired. It seems like not the amount but rather the quality
or fit of the patents could have a stronger impact on performance.
4.3 Summary of findings - multinominal logistic regression
Model Variables Dependent Significance Observations Calculation
1 d lcar21_1 2.810*** 1138 Periods
2 d lfreq_ma -0.424** 1615 Periods
3 d ltri_ar 2.510*** 1615 Periods
4 d lpi_ar 3.085*** 1615 Periods
5 d lcar21_1 3.027*** 680 Patents x Periods
6 d lfreq_ma, num_pat -0.440* 970 Patents x Periods
7 d ltri_ar, num_pat 2.841*** 970 Patents x Periods
8 d lpi_ar,num_pat 2.791*** 970 Patents x Periods
9 eup3 lcar21_1, num_pat -2.469** 680 Patents x Periods x Locations 10 eup4 lcar21_1, num_pat 1.238*** 637 Patents x Periods x Locations 11 eup4 ltri_ar, num_pat 2.181*** 865 Patents x Periods x Locations 12 eup4 lpi_ar,num_pat 2.121*** 970 Patents x Periods x Locations
13 ni lcar21_1 0.632*** 1138 Locations
14 ni ltri_ar 1.027*** 1615 Locations
15 ni lcar21_1, num_pat 0.556** 680 Patents x Locations 16 ni ltri_ar, num_pat 0.861*** 970 Patents x Locations
17 num_pat lfreq_ma 0.00208* 970 Patents
18 num_pat lfreq_ma, num_pat 0.00216** 970 Patents x Periods 19 num_pat lfreq_ma, num_pat 0.00206* 970 Patents x Locations 20 num_pat lfreq_ma, num_pat 0.00220** 970 Patents x Periods x Locations
21 p2 lcar21_1 -2.786*** 1138 Periods
22 p2 lfreq_ma -0.662*** 1615 Periods
23 p2 lpi_ar -3.514*** 1347 Periods
24 p2 lcar21_1 -2.137** 680 Patents x Periods
25 p2 lfreq_ma, num_pat -0.893*** 970 Patents x Periods 26 p2 lpi_ar,num_pat -3.048*** 970 Patents x Periods
27 p3 lcar21_1 -2.810*** 1138 Periods
28 p3 ltri_ar -1.792*** 1347 Periods
29 p3 lpi_ar -2.179*** 1268 Periods
30 p3 lcar21_1 -3.027*** 680 Patents x Periods
31 p3 ltri_ar, num_pat -2.127*** 970 Patents x Periods 32 p3 lpi_ar,num_pat -2.075*** 970 Patents x Periods
33 p4 lcar21_1 3.269*** 1138 Periods
34 p4 ltri_ar 3.610*** 1268 Periods
35 p4 lpi_ar 4.193*** 1615 Periods
36 p4 lcar21_1 3.363*** 680 Patents x Periods
37 p4 ltri_ar, num_pat 3.949*** 970 Patents x Periods
38 p4 lpi_ar,num_pat 3.889*** 970 Patents x Periods
39 row_tar ltri_ar 0.410* 1615 Locations
40 row_tar ltri_ar, num_pat 0.653** 970 Patents x Locations 41 rowp3 lcar21_1, num_pat -2.469** 680 Patents x Periods x Locations 42 rowp4 lcar21_1, num_pat 1.431*** 666 Patents x Periods x Locations 43 rowp4 ltri_ar, num_pat 2.006*** 941 Patents x Periods x Locations 44 rowp4 lpi_ar,num_pat 1.735*** 970 Patents x Periods x Locations
45 tp lpi_ar 0.401** 1615 Periods
46 u lcar21_1 -2.810*** 1138 Periods
48 u ltri_ar -2.510*** 1615 Periods
49 u lpi_ar -3.085*** 1615 Periods
50 u lcar21_1 -3.027*** 680 Patents x Periods
51 u lfreq_ma, num_pat 0.440* 970 Patents x Periods
52 u ltri_ar, num_pat -2.841*** 970 Patents x Periods
53 u lpi_ar,num_pat -2.791*** 970 Patents x Periods
54 us_tar lfreq_ma 0.287* 1615 Locations
55 us_tar ltri_ar -0.542** 1615 Locations
56 us_tar lfreq_ma, num_pat 0.375* 970 Patents x Locations 57 us_tar ltri_ar, num_pat -0.824*** 970 Patents x Locations 58 us_tar lpi_ar,num_pat -0.670** 970 Patents x Locations 59 usp1 lfreq_ma, num_pat 0.635** 970 Patents x Periods x Locations 60 usp2 lfreq_ma, num_pat -0.824* 970 Patents x Periods x Locations 61 usp2 lpi_ar,num_pat -2.407** 970 Patents x Periods x Locations 62 usp3 lcar21_1, num_pat -2.010*** 680 Patents x Periods x Locations 63 usp3 ltri_ar, num_pat -1.840** 970 Patents x Periods x Locations 64 usp3 lpi_ar,num_pat -2.509** 848 Patents x Periods x Locations 65 usp4 lcar21_1, num_pat 1.744*** 680 Patents x Periods x Locations 66 usp4 ltri_ar, num_pat 1.556*** 970 Patents x Periods x Locations 67 usp4 lpi_ar,num_pat 1.819*** 970 Patents x Periods x Locations
Those findings indicate that (regardless of location) acquiring companies in slight
economic downturns, indicated by period 4, has positive financial implications for the
acquirer. As mentioned in the results of the multivariate regression, slight economic
downturns could have benefits of getting cheaper deals or even benefitting of reduced risk
because a slight downturn can be seen as a temporary phase after which the market
regenerates for further growth. Further it can be interpreted that national M&A deals
account for positive abnormal returns because the transaction costs are the lowest when
integrating a company that functions in the same market and under the same legislation.
5. Discussion
This section is structured alongside the calculation models and discusses the implications
given by patents, periods and locations. The hypotheses will be answered in this section
accordingly.
5.1 Patents
The number of patents that have been traded does not influence the returns of the
company - therefore the innovativeness does not increase measurably.
No significance on number of patents on performance.
needed to distinguish what motives the US acquirers really had, and if they have relied to
heavy on the transaction cost theory.
The frequency of M&As influences the number of patents that have been traded.
All performed regressions in this study indicate a positive trend as it comes to the
frequency of M&As and the number of patents that have been traded. As it comes to this
research it is a logical implication that if the frequency of M&As goes up also the number of
patents that are transferred increases. Nevertheless, this result underlines the validity of
this approach to distinguish further, if there are related performance increases by financial
or growth indicators.
Regression Model Variables Dependent Sig. Obs. Calculation
mvreg 5 num_pat freq_ma, num_pat 0.00252* 511 Patents x Locations
mvreg 4 num_pat freq_ma 0.00251* 511 Patents
mvreg 6 num_pat freq_ma, num_pat 0.00250* 511 Patent x Periods x Locations
mlog 17 num_pat lfreq_ma 0.00208* 970 Patents
mlog 19 num_pat lfreq_ma, num_pat 0.00206* 970 Patents x Locations mlog 18 num_pat lfreq_ma, num_pat 0.00216** 970 Patents x Periods mlog 20 num_pat lfreq_ma, num_pat 0.00220** 970 Patents x Periods x Locations
For the above mentioned reasons hypothesis 1, stating that the number of patents that
have been transferred from the target to acquirer have a positive effect on innovative
performance, cannot be proven and has to be rejected.
The analysis shows no significance as it comes to the correlation of the number of patents
and returns. It is possible that this is an indicator of a rather strategic nature of acquiring
patents through M&As than for new product developments. The authors Artz et al. (2010)
found out in their longitudinal study of the impact of R&D, patents and product innovation
on firm performance, in which they investigated 272 companies in 35 industries, that
patents had a significantly negative relationship on ROA and sales growth. Their study
findings combined with the results of this research begs the question whether patents can
really indicate innovative performance as it comes to financial measures – or do firms use
patents rather as strategic weapons to secure their own position and block out
competitors. Further research would be needed to investigate to what extent traded
patents generated product innovations that can be commercialized, or whether those
product innovations come from internal R&D.
5.2 Periods
Pharmaceutical companies that acquire in economic downturns have higher positive
returns.
implications an economic downturn brings with it by reducing available funds and
increasing the uncertainty. My research is in line with their findings and adds another
insight, attributed to the turning points.
Avoiding turning points by +1/-1 Quartile when acquiring companies is essential to
not make negative returns.
The statistics show that peak performances, defined as turning points, should be avoided
by the timespan of approximately (+1 Quartile; -1 Quartile), otherwise the returns will
have a significant negative result. Therefore, concluding it can be said that not peak
performances of the market should be aimed at when performing mergers and
acquisitions, but rather slight economic downturns.
Regression Model Variables Dependent Sig. Obs. Calculation
reg 1 d freq_ma -0.223* 1614 Periods
mlog 5 d lcar21_1 3.027*** 680 Patents x Periods
mlog 6 d lfreq_ma, num_pat -0.440* 970 Patents x Periods
mlog 7 d ltri_ar, num_pat 2.841*** 970 Patents x Periods
mlog 8 d lpi_ar,num_pat 2.791*** 970 Patents x Periods
mlog 1 d lcar21_1 2.810*** 1138 Periods
mlog 2 d lfreq_ma -0.424** 1615 Periods
mlog 3 d ltri_ar 2.510*** 1615 Periods
mlog 4 d lpi_ar 3.085*** 1615 Periods
reg 17 tp car21_1 29.99* 1136 Periods
reg 18 tp tri_ar, pat -87.44*** 701 Patents x Periods
reg 19 tp tri_ar -59.96*** 1137 Periods
mvreg 15 tp tri_ar -77.94*** 827 Periods
mvreg 16 tp pi_ar -39.45* 827 Periods
mvreg 17 tp tri_ar, num_pat -100.0*** 511 Patents x Periods
mlog 45 tp lpi_ar 0.401** 1615 Periods
reg 20 u freq_ma 0.223* 1614 Periods
mvreg 18 u pi_ar 44.98** 827 Periods
mlog 50 u lcar21_1 -3.027*** 680 Patents x Periods
mlog 51 u lfreq_ma, num_pat 0.440* 970 Patents x Periods
mlog 52 u ltri_ar, num_pat -2.841*** 970 Patents x Periods
mlog 53 u lpi_ar,num_pat -2.791*** 970 Patents x Periods
mlog 46 u lcar21_1 -2.810*** 1138 Periods
mlog 47 u lfreq_ma 0.424** 1615 Periods
mlog 48 u ltri_ar -2.510*** 1615 Periods
mlog 49 u lpi_ar -3.085*** 1615 Periods
Figure 4 - Performance of Global Pharmaceutical Index 1997 -2008
The positive significance is supported by the results of the variable concerning the
abnormal growth percentage for the price index (pi_ar) for the global pharmaceutical
industry. Interestingly, the variable describing the abnormal return of growth percentage
of the total return index (tri_ar), shows a countercyclical trend for the turning points (see
Result section). The variable (tri_ar) therefore gives insights of the risk for the
shareholders, that even though the price index indicates a positive correlation for the
acquiring company itself, it fails to generate shareholder value in the long run.
The turning points in the analyses show significantly bad abnormal returns when it comes
to the creation of shareholder value and also in not increasing the stock price value of the
acquiring company. It seems like acquiring other companies in the timespan of (-1
Quartile; +1 Quartile) of the max / min performance of the global pharmaceutical index, is
negatively correlated to value creation for the company itself.
An explanation for this finding can be that the value of the deals that have been performed
in those periods were more expensive than in the other periods and drained therefore the
funds away from the company, lowering the shareholder’s equity.
When looking at figure 5, a pattern of high valued deals emerges, especially in period 2 and
3 and their turning points.
If we take a step further and take the inflation adjusted deal values and benchmark them
to the industry average the pattern intensifies (figure 6). It turns out that in the economic
upturn periods the deal values far exceeded the industry average and the turning points
indicate the highest peaks of spending habits.
Figure 6 – Inflation adjusted deal values benchmarked to industry average 1997 - 2008
Going further, the statistical analysis proves also that companies that acquire in economic
downturns have a significantly higher likelihood of turning out positive returns, than
companies that acquire in economic upturns. Therefore hypothesis 2, stating that
pharmaceutical companies that acquire in economic downturns have higher positive
returns, is proven.
5.3 Locations
The location of the target firms has an effect on the number of patents that have been
traded.
Throughout this research especially US based targets struck attention. The results indicate
that most of the target firms were US based and have been acquired by national deals.
Interesting however, is the bad financial performance these deals had as a consequence.
US based companies seem to have other motives to perform M&As than just to
increase their patent portfolio.
The US companies show a high activity as it comes to mergers, especially in period 2
defined as an economic downturn. Through their increased activity a significant number of
patents has also been transferred, which theoretically should have resulted in an increase
of the likelihood to generate innovations. The statistics show however, that quite the
contrary took place. While the US was keen to merge and acquire, the rest of the world was
rather passive, showed by a negative significance in period 2 as it comes to the frequency
of M&As. Accordingly, less patents were transferred and better financial abnormal returns
gained by the company and shareholders, in the long- and short-term.
Regression Model Variables Dependent Sig. Obs. Calculation
mlog 12 eup4 lpi_ar,num_pat 2.121*** 970 Patents x Periods x Locations
mlog 13 ni lcar21_1 0.632*** 1138 Locations
mlog 14 ni ltri_ar 1.027*** 1615 Locations
mlog 15 ni lcar21_1, num_pat 0.556** 680 Patents x Locations mlog 16 ni ltri_ar, num_pat 0.861*** 970 Patents x Locations mvreg 11 row_tar car21_1, num_pat 50.66* 511 Patents x Locations mvreg 12 row_tar pi_ar, num_pat 58.01** 511 Patents x Locations
mlog 39 row_tar ltri_ar 0.410* 1615 Locations
mlog 40 row_tar ltri_ar, num_pat 0.653** 970 Patents x Locations reg 16 rowp3 pi_ar, num_pat 58.49* 701 Patents x Periods x Locations mvreg 13 rowp3 car21_1, num_pat 71.18* 511 Patent x Periods x Locations mvreg 14 rowp3 pi_ar, num_pat 71.02** 511 Patent x Periods x Locations mlog 41 rowp3 lcar21_1, num_pat -2.469** 680 Patents x Periods x Locations mlog 42 rowp4 lcar21_1, num_pat 1.431*** 666 Patents x Periods x Locations mlog 43 rowp4 ltri_ar, num_pat 2.006*** 941 Patents x Periods x Locations mlog 44 rowp4 lpi_ar,num_pat 1.735*** 970 Patents x Periods x Locations
mlog 54 us_tar lfreq_ma 0.287* 1615 Locations
mlog 55 us_tar ltri_ar -0.542** 1615 Locations
mlog 56 us_tar lfreq_ma, num_pat 0.375* 970 Patents x Locations mlog 57 us_tar ltri_ar, num_pat -0.824*** 970 Patents x Locations mlog 58 us_tar lpi_ar,num_pat -0.670** 970 Patents x Locations mlog 62 usp3 lcar21_1, num_pat -2.010*** 680 Patents x Periods x Locations mlog 63 usp3 ltri_ar, num_pat -1.840** 970 Patents x Periods x Locations mlog 64 usp3 lpi_ar,num_pat -2.509** 848 Patents x Periods x Locations reg 24 usp4 car21_1, num_pat 0.549* 970 Patents x Periods x Locations mlog 65 usp4 lcar21_1, num_pat 1.744*** 680 Patents x Periods x Locations mlog 66 usp4 ltri_ar, num_pat 1.556*** 970 Patents x Periods x Locations mlog 67 usp4 lpi_ar,num_pat 1.819*** 970 Patents x Periods x Locations
