☆ We thank Håkan Jankensgård, our supervisor, for providing insightful comments. We would also like to thank Bengt EW Nilsson for fruitful discussions and Daniel Tano for proofreading and revising. Finally, we would like to thank Department of Economics and Management. a* Lund University, MSc in Finance Graduate, [email protected], 881124-4071 b* Lund University, MSc in Finance Graduate, [email protected], 890125-4832 INFO ABSTRACT Corporate inversion is the procedure of reincorporating abroad for tax purposes to reduce the corporate income tax. The purpose of this paper is two-fold. First, we evaluate the shareholder value creation aspect of U.S. corporate inversions after 2004, a year when the U.S. imposed a number of regulatory changes to reduce the benefits of tax inversions. Second, we examine a number of board- related determinants of the decision to invert, with particular focus on board- level patriotism and firm ethics. By using traditional event study methodology we show that inversions create significant shareholder value. We also find that the ratio of foreign professionals on U.S. corporate boards positively affect the decision to invert. By assuming a positive relationship between patriotism and the level of board reputational costs, we introduce the notion that lack of board- level patriotism may partly explain the under sheltering problem related to tax inversions. Corporate Tax Inversions and Reputational Costs ☆ - A Study on Board-Level Patriotism and the Decision to Expatriate Alexander Gjörup a* Gustav Nilsson b* School of Economics and Management, Lund University, Sweden 25 May 2016 Available Online: 19 June 2016 Keywords: Tax Avoidance Tax Inversion Board Nationality Reputational Costs Under Sheltering Puzzle
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
☆ We thank Håkan Jankensgård, our supervisor, for
providing insightful comments. We would also like to thank
Bengt EW Nilsson for fruitful discussions and Daniel Tano
for proofreading and revising. Finally, we would like to
Shulman (2009), for instance, articulated in a speech that aggressive tax avoidance imposes a
―significant risk to corporate reputations‖ and that ―the general public has little tolerance for
overly aggressive tax planning‖.
13
This leads us to our first hypothesis; as prior research has shown that ―ordinary‖ M&A
activity does not create any significant value for the acquiring firm, we hypothesize that
adding the extra benefit of potential tax savings that come with inversions should increase
shareholder value. In other words, we wonder if the benefits generated from tax inversions are
enough to off-set the tax and non-tax related costs associated with such transactions.
Hypothesis 1: For acquiring firms, tax inversions create shareholder value.
The following two sections are structured as follows. First, an overview of previous
findings regarding firm-determinants of tax inverting firms is laid out. Second, we present a
new set of board-related determinants which we argue are central in this debate and therefore
deserve greater attention in both the academic community. Hence, by studying factors that
directly impact corporate leaders‘ decision making, we hope to shed new light on various
aspects impacting firms‘ propensity to invert.
3.3 Firm-specific determinants of tax inverting firms
Although focused on the time prior to the 2004 American Job Act17
, as mentioned in
section II (regulatory background), Desai & Hines (2002) brings up a number of firm specific
determinants of tax inverting firms. They present evidence that larger firms with extensive
foreign assets, along with significant leverage, are more likely to invert. Similar findings are
found by Col, Liao & Zeume (2016), who present evidence that inverting firms are often
profitable, highly leveraged and earn significant amount of their revenues abroad.
Significant cash holdings of U.S. firms with international subsidiaries are another potential
determinant of firms deciding to re-domicile through corporate inversions. Foley, Hartzell,
Titman & Twite (2007), for instance, show that U.S. multinationals hold significant part of
their assets in cash due to the significant tax costs associated with the repatriation of cash
back to the U.S.18
Col, Liao, & Zeume (2016) also find a positive relation between cash
position and propensity to invert. Similar findings regarding higher cash levels are found by
17
2004 American Job Act as described in section II (Regulatory background). N.B. earlier work published before 2004 in
this field of research should be deemed less relevant due to ongoing changes in the regulatory environment; hence, work
published before the major regulatory American Job Act form 2004 has been given limited exposure in this paper. Another
aspect of older research published in this area is that due to regulatory changes different type of inversions have been
analyzed in different periods. For instance, as most tax haven inversions (see regulatory section) disappeared after the 2004
AJC, we consider those inversions less relevant in today‘s debate. 18 Refer back to section II (Regulatory background) for more information on repatriation of cash to the U.S.
14
Cortes, Gomes & Gopalan (2014), who presents findings that American companies in foreign
countries tend to have more cash on hand than do U.S. based firms.
Another important aspect recently discussed is whether compensation and other
incentivizing factors may drive corporate leaders' propensity to carry out inversions. Babkin,
Glover & Levin (2015) find that although returns for private individuals are negative,
particularly those with medium and long-term investment horizons, the returns for CEOs as a
result of such transactions tend to be positive. This is, they argue, due to the stock option
nature of CEO compensation programs, which are taxed favorably compared to private
individuals‘ capital gains (Babkin, Glover & Levine, 2015). On top of the beneficial tax
schedule for capital gains on options, they find that CEOs of inverting firms receive on
average an additional $3.26 million above their normal pay the year following a tax inversion
(Babkin, Glover & Levine, 2015).
Industry specification has also been analyzed to be one of the major determinants of tax
inverting firms due to the way certain tax heavens treat intangible assets, which, as an
example, constitutes the majority of the assets of pharmaceutical companies (Sharife, 2016).
Bringing up drug-giant Pfizer as an example, in 2012, the ca. $190 billion market19
capitalization company had net tangible assets (NTA) just shy of $6 billion (Sharife, 2016).
Couple that with the fact that companies in the Netherlands, for instance, face a mere 5% tax
rate on revenues stemming from certain intangible assets, such as patents and other R&D
related activities (Nijhof & Kloes, 2010). Desai & Dharmapala (2006) also finds evidence
supporting the notion that high R&D intensive firms are more likely to invert into tax heaven
countries than are other firms.
In additional to firm-specific determinants of tax inverting firms, previous research has
also explored other perhaps less obvious potential determinants. One such paper, published by
Col, Laio & Zeume (2016), seeks to go beyond firm specific determinants and focus more on
country-related corporate governance aspects of tax inversions. Using a large sample of
inverting firms globally, they find that firms are more likely to invert to countries with strong
governance standards and that country-pair specific agreements, such as bilateral Double
Taxation Treaties (DTTs)20
and Tax Information Exchange Agreements (TIEAs)21
, are likely
to increase the amount of inversions.
19 As of 2016-05-09, Pfizer‘s market capitalization was $207.9 billion. 20 DTTs are explained further by Col, Laio & Zeume (2016) 21 TIEAs are explained further by Col, Laio & Zeume (2016)
15
3.4 Board-related determinants of tax inverting firms
Despite the increasing amount of research being published on the topic of tax inversions,
as summarized in previous sections - often centered around shareholder returns or value
creation, along with firm-specific determinants - few conclusions have been drawn around
board-related determinants of tax inverting firms. Therefore, as a way to tackle this gap in the
literature, our objective is to further analyze why firms do not engage more in tax inversions
although there are a number of clear financial benefits associated with them as laid out in the
previous sections.
Foreign directors
In connecting the very limited literature regarding corporate governance and patriotism,
legal scholars such as Yosifon (2015), argue that board of directors‘ decision making may be
impacted by their level of patriotism and that it is reasonable to discuss the consequences of
corporate board members being of different nationalities. If ―patriotism is a virtue‖, he argues,
then U.S. firms should hire primarily American leaders in order to honor American interests.
Some earlier legal research point towards the same direction, putting forward the importance
of the trade-off between shareholder value creation and patriotic behavior. That is, since
firms‘ ultimate objective is to increase shareholder value (Mann, 2004), which in today‘s
globalized economy may require American companies to move both jobs and tax revenues
abroad (Yosifon, 2015), then companies must ignore patriotic objectives in order to increase
shareholder value and to remain competitive. Kate Barton, for instance, who is a partner of
Ernst and Young states that ―the improvement on earnings is powerful enough that maybe the
patriotism issue need to take a back seat‖ (Mann, 2004).
We are first in attempting to empirically connect patriotism with firms‘ decision to invert.
However, some related research does exist and is important to put forward in order to
understand the full scope of this debate. For instance, previous research show that in certain
cases foreign board members have a positive net impact on firm value, particularly for firms
active in cross-border M&A transactions (Masulis, Wang & Xie 2012). The authors argue that
foreign board members have better knowledge of local markets, leading to increased ability to
evaluate potential targets. Minnick and Noga (2010) apply a similar reasoning while arguing
that different tax planning strategies may require different skillsets. For instance, international
tax planning strategies, they argue, are more complex in nature and therefore require more
specialized expertise. Similarly, in a study bringing up the costs and benefits of foreign
directors, where 80 international (non-U.S.) firms were studied, Miletkov & Poulsen (2013),
16
show that companies engaging in cross-border M&A transactions are more likely to have
foreign directors on their boards.
On the cost side of the equation, Miletkov & Poulsen (2013) proposes the idea that foreign
board members may impose organizational challenges in terms of less effective decision
making and worse monitoring of management. Banerjee, Masulis & Pal (2016), for instance,
show that foreign board members are significantly more likely to miss board meetings than
are American board members, suggesting weaker ability to control and impact board
decisions, as well as monitoring of corporate leaders.
In all, based on previous rather limited empirical research, it seems as investors appreciate
the existence of foreign board members on U.S. corporate boards, at least when it relates to
firms highly active in cross-border M&A transactions. Our objective, however, is more
closely in line with the discussion brought to attention by legal scholars (e.g Yosifon, 2015)
that American companies are increasingly putting aside patriotic objectives for shareholder
value creation. More specifically, as a way of understanding the importance of patriotism, we
are interested in determining whether U.S. corporate boards with higher ratio of foreign board
members are more likely to engage in tax inversions. Inspired by Gallamore, Maydew &
Thornock (2013)22
, is it perhaps that foreigners on U.S. corporate boards by definition lower
overall firm-wide patriotism, leading to lower inherent reputational costs for such firms,
thereby increasing firms‘ propensity to invert23
.
Hypothesis 2a: High rates of non-U.S. board members increase firms‘ propensity to invert
Corporate Social Responsibility
The academic consensus regarding CSR, ethics and corporate strategic decision making
seem to point towards a reality where social responsibility and corporate actions go hand in
hand (Goda, O‘Connor & Taylor 2005; Joyner & Payne, 2012; Garriga & Mele, 2003;
Godfrey & Merrill, 2009). For instance, Joyner and Payne (2012) argue that it is a
―fundamental truth‖ that firms and the society cannot survive without each other; hence, firms
must comply with society‘s increased demand of ethical behavior and social responsibility to
22
Although finding no evidence for such reasoning, Gallamore, Maydew, Thomock (2013) hypothesize that only
firms that are immune to reputational costs engage in tax avoidance activities. 23
N.B. we are cautious regarding the by us assumed relationship between patriotism and reputational costs. That
is, we have no empirical evidence that higher degree of patriotism leads directly to higher reputational costs;
instead, this is an assumption based on intuitive reasoning.
17
remain competitive (Joyne and Payne, 2012). Related empirical literature discusses similar
issues. For instance, Lanis & Richardson (2012) show that firms with high CSR engagements
are likely to be less tax aggressive. From a slightly different angle, Godfrey & Merrill (2009)
states that CSR engagement is a way to convey the society that firms are not acting in pure
self-interest, but rather to contribute positively to the society and its various stakeholders.
Therefore, naturally, extensive tax planning cannot easily be reconciled with ethical business
behavior (Godfrey & Merrill 2009). In fact, Sikka (2010) find that many firms act
hypocritically when dealing with CSR related activities; that is, firms engage in CSR related
activities while they are at the same time actively pursuing various tax planning strategies,
including moving assets to foreign tax heavens.
Furthermore, based on one of the first articles (Huseynov & Klamm, 2012) regarding the
relationship between tax avoidance and CSR engagement, the authors find that there is a
relationship between the two variables and conclude that firms with strong community
concerns pay higher effective tax rates, while corporate governance strength and firm
diversity is negatively related to effective rax rates. In other words, certain dimensions of
CSR may have one impact on a firm‘s propensity to pay taxes and engage in tax planning,
while other dimensions of CSR may have another impact firms‘ engagement in tax planning
activities.
Hence, our second hypothesis is based on the more traditional notion that increased CSR
should lead to more engagement in activities that support the society as a whole and the
greater good of the area in which it operates and less in activities, such as tax inversions, that
do not contribute to the society overall24
.
Hypothesis 2b: High degrees of CSR engagement decrease firms‘ propensity to invert
Entrenched Boards
To study the effect of entrenchment on firms‘ propensity to engage in tax inversions, we
study whether firms with staggered boards are more likely to invert. As described by Adams,
Hermalin and Weisbach (2008), staggered board structures are such that board directors are
elected for multiple years at a time, where only a fraction of the board is re-elected each year.
This creates a situation where it becomes difficult for outside investors to impact the
24
In contrary to the common notion and consensus in the public debate a recent study by Gunn & Lys (2015)
show that tax inversions are actually, paradoxically, likely to add value for the U.S. Treasury Department
through increased tax revenues.
18
composition of the board, thereby reducing the risk of losing control of the firm (Bebchuk &
Cohen, 2005; Cohen & Wang, 2013), e.g. through stand-alone proxy contests25
and hostile
takeovers26
.
Past literature show evidence that the existence of entrenchment decreases the likelihood
that firms engage in tax avoidance related activities. Yijiang & Kung (2008), for instance,
present evidence that managers in firms with staggered boards are less motivated to maximize
profits to increase firm value, instead staggered board structures allow managers to live the
quite life (Bertrand & Mullainathan, 2003). This leads managers to be less likely to engage in
various types of tax avoidance activities (Minnick & Noga, 2010). More specifically, Minnick
and Noga (2010) find evidence that firms with entrenched boards pay more domestic taxes
than do their non-entrenched peers, suggesting that entrenched boards are less interested in
tax planning activities. This is consistent with findings presented by Bebchuk & Cohen (2005)
who argues that increased entrenchment lowers firm value. That is, entrenched managers are
not always acting in the best interest of shareholders.
Hence, our third hypothesis is grounded around the notion that entrenched managers are
less likely to engage in tax avoidance related activities, such as tax inversions, due to the idea
that entrenchment leads managers to be less interested in managing taxes.
Hypothesis 2c: High degrees of entrenchment decrease firms‘ propensity to invert
Insider Ownership
The fundamental connection between insider ownership and corporate tax avoidance
hinges on the notion that aligned interest between managers and shareholder interests should
increase a firm‘s propensity to engage in tax planning activities (Desai and Dharmapala,
2006), e.g. tax inversions. Such alignment often strengthens as a consequence of greater
incentives for management, e.g. through equity ownership. A similar discussion is brought to
attention by Armstrong, Blouin, Jagolinzer & Larcker (2013). They argue that CEO and
management equity incentives are related to firms‘ propensity to engage in tax avoidance
activities. More specifically, they suggest that tax avoidance activities shall be seen as
investments, where managers with higher equity incentives are more likely to engage in such
activities.
25
Proxy contest; an action by activists or other shareholders to use their voting power to impact the board or
management composition. 26
Hostile takeover; an action by a firm‘s board and/or management to convince a target firm‘s shareholders to
accept a takeover bid.
19
With findings from previous literature on insider ownership and general tax avoidance
activities in mind, we propose that our sample of tax inversion transactions should yield
similar findings that increased insider ownership increases firms‘ propensity to invert.
Hypothesis 2d: High degrees of insider ownership increase firms‘ propensity to invert
Institutional Ownership
In the literature it is common to connect institutional ownership and corporate governance
based on the notion that institutions with high ownership are more likely to monitor and
control corporate decision making (Ogden, 2003). Desai and Dharmapala (2006), for instance,
use institutional ownership as part of a proxy for institutional ownership, to understand firms‘
propensity to engage in various types of tax sheltering activities.
Nevertheless, the academic community is rather inconclusive as to what impact
institutional ownership may have on tax avoidance related activities; parts of the academic
community argue that institutional ownership is positively related to firms‘ propensity to
engage in tax avoidance related activities and others argue the opposite. Bird and Karolyi
(2015) support the former notion. They find that a 1 percent increase in institutional
ownership increases firms‘ propensity to have a subsidiary in a tax haven by 1.3 percent.
Thus, they suggest that institutional ownership increases the likelihood of firms engaging in
various tax planning activities. Similar findings are presented by Babkin, Glover & Levine
(2015), who find that institutional ownership increases firms‘ propensity to invert. On the
contrary, Khurana and Moser (2012) support the latter alternative. When analyzing corporate
data from 1995-2008 the authors find evidence that institutional ownership, particularly if
long-term oriented, tend to lower firms‘ likelihood to engage in tax avoidance.
Hence, our fifth and last hypothesis is based on the implications of external monitoring on
firms‘ decision making and how that relates to firms‘ propensity to invert. We hypothesis that,
and in line with findings founds by Bird & Karolyi (2015) and Babkin, Glover & Levine
(2015), increased institutional ownership increases firms‘ propensity to invert since
institutional investors, just like managers with high equity stakes, are likely to act in such way
that supports the share price. While our hypothesis is similar to the one provided by Babkin,
Glover & Levine (2015), we contribute by having a more homogenous transaction sample.
Hypothesis 2e: High degrees of institutional ownership increase firms‘ propensity to invert
20
3.5 Summary of hypotheses
Our first hypothesis is analyzed using a common event study methodology often used in
more traditional M&A studies. We study two samples of M&A transactions, where our main
sample of interest is focused around merger inversion transactions post-2004 and our control
sample includes similar transactions not labeled as tax inversions. In matching transactions
one by one we are able to control for firm specific determinants known to drive abnormal
returns, such as size of the acquirer and industry type. We also match announcement days by
year to control for time-varying effects. In using evidence from previous literature, primarily
related to the regulation of tax inversions, as well the financial benefits and potential costs of
such transactions, we construct the following hypothesis:
Hypothesis 1 For acquiring firms, tax inversions create shareholder value.
The second hypothesis consists of five different sub-categories which are all based on the
notion that certain board-related determinants should help us understand board decision-
making as it relates to tax inversions. We employ a logit regression to determine such
relationships, much inspired by similar studies in the field (e.g. Babkin, Glover & Levine,
2015).
Hypothesis 2a High rates of non-U.S. board members increase firms‘ propensity to invert
Hypothesis 2b High degrees of CSR engagement decrease firms‘ propensity to invert
Hypothesis 2c High degrees of entrenchment decrease firms‘ propensity to invert
Hypothesis 2d High degrees of insider-ownership increase firms‘ propensity to invert
Hypothesis 2e High degrees of institutional ownership increase firms‘ propensity to invert
IV. Methodology
This chapter is structured as follows; first, we walk through our data set and describe the
two samples and how they are collected. Secondly, we provide a summary of the event study
methodology used in this study. We conclude by presenting our cross-sectional regressions
used to analyze whether the two samples‘ CAAR are statistically different from each other as
well as our cross sectional logit regressions used to identify the impact of various board-
related determinants on firms‘ propensity to invert.
21
4.1 Data collection
In this study we employ a unique data set containing all, to our knowledge, announced tax
inversion transactions post-2004. The reason for choosing only transactions after 2004 is
because of the Americans Job Act regulation (see Section II; Regulatory background) that
changed the ecosystem of tax inversions. Our sample is similar to Gunn and Lys (2016), but
different from Babkin, Glover & Levine (2015) along with earlier studies from the early
2000‘s. In contrary to the sample used by Gun and Lys (2015), however, we only use merger
inversions, while they include re-incorporations and spin/split-offs; hence the difference in
sample sizes. After screening for post-2004 transactions, as well as removing non-merger
inversions, a final sample of 28 transactions is identified. Two additional transactions are re-
moved due to overlapping event windows (McKinley, 1989).
Using a larger sample size would have the possibility to give better validity and
significance of our results. However, our decision to only include a certain type of
transactions increases our ability to draw correct inference around the determinants of merger
inversions, specifically. This decision enables us to compare post-2004 merger inversions,
often directed to European countries with high governance laws with pre-2004 re-
incorporation inversions often directed to countries with low governance laws. Hence, by only
studying merger inversions we focus our research towards one type of transactions, which by
itself contributes to the current rather limited literature on tax inversions.
The data is collected from five different databases, including Merger Market, Thomson
Reuters Datastream & Eikon, Bureau van Djik‘s Zephyr, SEC Edgar and Capital IQ. For full
description of how the different databases27
are used to specify the inversion sample28
of 26
transactions and the peer group sample29
of 26 transactions visit appendix 1-3. We construct
the peer sample by matching each one of the tax inversions with transactions where U.S. firms
acquire an international firm but where the country of domicile remains in the U.S.
Announcement dates are used as the sole confirmation as to whether the firm chose to
invert or not, i.e., we do not account for what actually happened after the announcement.
Sample correction due to event clustering
27
See Appendix A1 28
See Appendix A2 29
See Appendix A3
22
The analysis of aggregating abnormal returns (AAR) assumes that event windows of the
included transactions do not overlap. This assumption is crucial when calculating the variance
of the AAR without having concerns about the covariance across returns (MacKinlay, 1997).
In order to avoid overlapping, often called event clustering, two transactions are removed.
Hence, the original sample of 28 transactions is reduced to 26.
In addition to the two transactions, we remove two other transactions that overlap by three
days. However, due to our small sample sizes these transactions remain in our analysis. In
fact, studies have shown that partial overlaps for shorter event windows (≤11 days) do not
produce any measurable bias (Karafiath, 2008).
23
Company Announced Target Destination Industry Market Value* Description***
Actavis 20-05-2013 Warner Chilcott Ireland Pharmaceuticals 88,9B Actavis Inc, a specialty pharmaceutical company, develops, manufactures, markets, and distributes
medical aesthetics, biosimilar, and over-the-counter pharmaceutical products worldwide.
Alkermes** 09-05-2011 Elan Drag Technologies Ireland Pharmaceuticals 5,99B Alkermes plc, a biopharmaceutical company, researches, develops, and commercializes pharmaceutical
products that are designed to address unmet medical needs of patients in various therapeutic areas
worldwide
Applied Materials Inc. 24-09-2013 Tokyo Electron Netherlands Industrials 22,59B Applied Materials, Inc. provides manufacturing equipment, services, and software to the semiconductor,
display, solar photovoltaic (PV), and related industries worldwide.
Argonaut 14-03-2007 PXRE Bermuda Insurrance 446,74M Argonaut Gold Inc. engages in the exploration, development, and production of gold and silver in North
America.
Arris Group 22-04-2015 Unit of OCI NV UK Telecom 4,42B ARRIS International plc provides media entertainment and data communications solutions in the United
States and internationally.
Burger King (RBI) 26-08-2014 Tim Hortons Canada Food 9,94B Restaurant Brands International Inc. owns and operates quick service restaurants under the Burger King
and Tim Hortons brand names.
C&J Energy Services Ltd. 25-06-2014 Nabor Industries Ltd. Bermuda Energy 125,02M C&J Energy Services, Ltd. provides completion and production services for oil and gas industry
primarily in North America.
Chiquita Brands 10-03-2014 Fyffes Ireland Food 668,65M Chiquita Brands International, Inc., together with its subsidiaries, markets and distributes bananas,
pineapples, and packaged salads primarily in the United States, Central America, Europe, the Middle
East, and Asia.
Eaton 21-05-2012 Cooper Industries Ireland Energy 28,44B Eaton Corporation plc operates as a power management company worldwide.
Endo International 05-11-2013 Paladin Labs Ireland Pharmaceuticals 3,46B Endo International plc develops, manufactures, and distributes pharmaceutical products and devices
worldwide.
Horizon Pharma 19-03-2014 Vidara Technologies Ireland Pharmaceuticals 2,35B Horizon Pharma plc, a biopharmaceutical company, engages in identifying, developing, acquiring, and
commercializing medicines for the treatment of arthritis, pain, inflammatory, and/or orphan diseases in the
United States and internationally.
IHS Inc 21-03-2016 Markit Inc UK Data 8,22B HS Inc. provides information, insights, and analytics to multinational companies, governments, small
companies, and technical professionals in various industries worldwide.
Jazz Pharmaceuticals 19-09-2011 Azur Pharma Ireland Pharmaceuticals 8,93B Jazz Pharmaceuticals Public Limited Company, a biopharmaceutical company, identifies, develops, and
commercializes pharmaceutical products for various medical needs in the United States, Europe, and
internationally
Johnson Controls 25-01-2016 Tyco International Ireland Industrials 26,7B Johnson Controls, Inc. operates as a diversified technology and industrial company worldwide.
Liberty Global 05-02-2013 Virgin Media UK Telecom 31,55B Liberty Global plc, together with its subsidiaries, provides video, broadband Internet, fixed-line
telephony, and mobile services in Europe, Chile, and Puerto Rico
Livanova (Cyberonics) 26-02-2015 Sorin UK Pharmaceuticals 2,34B LivaNova PLC, a medical technology company, design, develops, manufactures, and sells therapeutic
solutions worldwide.
Medtronic 15-06-2014 Coviden Ireland Pharmaceuticals 113,81B Medtronic plc manufactures and sells device-based medical therapies worldwide.
Mylan 14-08-2014 Abbott's generics unit Netherlands Pharmaceuticals 20,75B Mylan N.V., together with its subsidiaries, develops, licenses, manufactures, markets, and distributes
generic, branded generic, and specialty pharmaceuticals worldwide.
Pentair Ltd. 28-09-2012 Tyco Flow Control LTD Ireland Industrials 10,54B Pentair plc operates as a diversified industrial manufacturing company in the United States, Europe, and
internationally.
Perrigo 29-07-2013 Elan Ireland Pharmaceuticals 13,33B Perrigo Company plc, together with its subsidiaries, develops, manufactures, markets, and distributes
over-the-counter (OTC) consumer goods and pharmaceutical products worldwide.
Pfizer 23-11-2015 Allegran Ireland Pharmaceuticals 221.7B Pfizer Inc., a biopharmaceutical company, discovers, develops, manufactures, and sells healthcare
products worldwide.
Steris 13-10-2014 Synergy Health UK Pharmaceuticals 6,23B Steris Plc provides infection prevention, contamination control, surgical, and critical care technologies
worldwide.
Stratasys 16-04-2012 Objet Israel Data 1,13B Stratasys Ltd. provides three-dimensional (3D) printing and additive manufacturing (AM) solutions for
the creation of parts used in the processes of designing and manufacturing products.
Terex 11-08-2015 Konecranes Finland Construction 2,69B Terex Corporation operates as a lifting and material handling solutions company
Tower Group 05-08-2008 Castlepoint Holdings Bermuda Insurance 127,77M Tower Group provides specialized property and casualty insurance products and services to small to
mid-sized businesses and to individuals in New York City and the adjacent areas of New York State.
Wright Medical 27-10-2014 Tornier Netherlands Pharmaceuticals 1,9B Wright Medical Group N.V., a medical device company, designs, manufactures, markets, and sells
orthopedic products in the United States, Europe, and internationally.
** Alkermes Presents Phase 2 Data of ALKS 37 in Late-Breaking Oral Session at Digestive Disease Week Meeting at Announcement Day of Inversion.
*** Obtained from Mergermarket
The final sample of included tax inversions with accompanied transaction data.
List of Tax Inversions
* Market value of equity is measured at the announcement date.
24
4.2 Event Study
We use a standard event study methodology proposed by MacKinlay (1997) to examine
the market reaction around the announcement day of tax inversions. We start by defining an
event window, during which the actual event occurs, and an estimation window, during which
our asset model parameters are estimated. We use our estimated parameters to estimate the
expected normal returns in the event window. Lastly, abnormal returns, AR, are calculated by
the difference between the actual return on the stock during the event window and the
expected normal return for the company if no event had occurred. Regarding the literature on
tax inversions, event studies are relatively common (e.g. Desai & Hines, 2002; Seida &
Wempe, 2002; Babkin, Glover & Levine, 2015).
As proposed by MacKinlay (1997), our estimation window consists of 120 trading days,
while our event windows consist of two, three, seven and eleven days. In order to properly
estimate the model parameters used to calculate the expected normal returns, the event
window is set to start the day after the estimation window ends. This is important because the
parameters used in the event study should not be influenced by the possible abnormal returns
caused by the event (MacKinlay 1997).
The event day is the day during which our sample firms for the first time announce their
intention to invert. These dates are identified in Merger Market and through EDGAR filings.
In cases where announcements were made when the market was not open, e.g. at night or on
the weekends, the first day of trading after the announcement day is used as the event day.
Using the announcement day, T, as event the event day has been shown suitable when
measuring wealth effects of M&A activity (Datta et al 1992).
𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑖𝑜𝑛 𝑤𝑖𝑛𝑑𝑜𝑤 𝑜𝑓 120 𝑡𝑟𝑎𝑑𝑖𝑛𝑔 𝑑𝑎𝑦
𝐸𝑣𝑒𝑛𝑡 𝑤𝑖𝑛𝑑𝑜𝑤 𝑑 𝜖 (1,2,3,5)
𝑇
𝑇 + 𝑑
𝑇 − 𝑑
𝑇;125
25
Estimating normal returns using the market model
There are a number of different approaches to estimate the expected normal returns for
equities, including the Constant Mean Return Model and the Market Model, both of which are
statistical models based on various statistical assumptions (MacKinley, 1997). In this study
the market model is used as a way to estimate the normal expected returns during the event
window. The statistical assumptions underlying this model are that asset returns are jointly
multivariate and independently and identically distributed (IID) (MacKinlay 1997). These
assumptions are sufficient for the Market Model to correctly estimate our model parameters
(MacKinlay 1997).
Further, we use the market model since it has been frequently used when evaluating the
stock price reactions to announcements of M&A transactions (Shleifer, A., & Vishny, R. W.
(2003); Fama, Fisher, Jensen & Roll, 1969). The market model is also relatively easy and
straight forward to use, while at the same time generating proper and sufficient parameter
estimates. Lastly, and compared to other models, such as the Constant Mean Return model,
MacKinley (1997) argues that the Market Model provides superior parameter estimates with
lower variance, which increases the ability to capture abnormal returns. In this study, we use
the same notations as proposed by MacKinlay (1997) when presenting our Market Model
formula.
+ + 1.
( 0) ( 2) 2.
OLS regressions of on are used to estimate the Market Model parameters
(intercept, 30 and the slope,
31), where is the return of stock, i, and is the return
of the market portfolio. Since all our sample firms are U.S. companies, we use the S&P500
Composite to reflect the returns of the market portfolio. This index is the most commonly
used and well-known U.S. stock index covering the 500 largest stocks in different industries.
The returns for the S&P500 Composite are retrieved through Thomson Reuters Datastream
using time series requests and the symbol S&PCOMP. Given the estimated parameters the
expected returns32
during the event windows are calculated.
30
−
31
∑ (
; )( ; )
∑ (
; )
32 +
26
Abnormal Returns
We estimate the firm-specific abnormal returns for each day within the event window
using following formula:
− 3.
Where is the expected normal return for stock, i, at time t. Next we aggregate the
abnormal returns across firms for each day to obtain sample-wide average abnormal returns,
or AAR.
1
∑
<1
4.
Cumulative Abnormal Returns
Stock returns are also aggregated across time in order analyze multi-period event windows.
Formula (5) visualizes how we aggregate AARs through time to retrieve a sample-specific
cumulative average abnormal return (CAAR), or
( ; , : ) ∑
:
< ;
Alternatively, as shown in formula (6), one could also take the average of the
estimated firm-specific cumulative abnormal returns (CAR) to retrieve the same
result as in formula (5).
5.
( ; , : ) 1
∑ ( ; , : )
:
< ;
6.
As a way to verifying the abnormal returns we present p-values indicating the significance
of our results, as well adjusting the sample returns for outliers.
27
4.3 Modeling the market reaction of inversion
To draw further theoretical insights from our results in the event study we apply several
cross-sectional regression models according to the principle of MacKinlay (1997). We use
these models to investigate how different firm and transaction specific variables affect returns.
Most importantly, we run these regressions with the objective to evaluate how CARs are
affected by transactions characterized as tax inversions, which would help us verify the
descriptive results from the initial event study. In the model, a dummy variable is used to
denote whether the transaction is a tax inversion or a regular international cross-border M&A.
In addition, several control variables are included to control for other factors that have already
been proven to drive shareholder value in M&A transactions (see variables list).
The first regression is related to our first hypothesis (section 3.5; Summary of hypotheses);
that is to gauge whether tax inversions generate higher, lower or similar CAAR than do non-
inversion transactions. In order to examine such relationships, we run both a univariate and
multivariate regression for several event windows. In essence, the multivariate regression is
an extended version of the univariate model that tries to isolate the effect of our tax inversions
on CAAR by controlling for other already proven drivers of CAAR.
The univariate and multivariate regressions are specified in (7 & 8). We run them for all
our estimated event windows. For fully specified regressions and list of variables (see