Economic Modelling xxx (xxxx) xxx

Contents lists available at ScienceDirect

Economic Modelling
journal homepage: www.journals.elsevier.com/economic-modelling

Does eliminating international profit shifting increase tax revenue in

high-tax countries?
Patrice Pieretti 1 , Giuseppe Pulina 1, ∗
University of Luxembourg, Department of Economics and Management, Luxembourg

A R T I C L E I N F O A B S T R A C T

JEL classification: We analyze under what conditions initiatives intended to eliminate profit shifting (such as the OECD BEPS
F21 action plan and the more recent European implementation of the ATAD) can be successful, given that these
F23 actions may induce multinational companies to relocate activities to low-tax countries. We demonstrate that
H25
removing tax-motivated profit shifting increases tax revenue in the onshore region if the low-tax jurisdiction is
H26
not too efficient in providing attractive infrastructure. This outcome is more easily achieved when the high-tax
country is able to counter the shifting of international activity using infrastructure investment to compete with
Keywords: the tax haven rather than being passive. International regulations aimed at combating aggressive tax avoidance
BEPS
should anticipate the adverse effects induced by the resulting emergence of other forms of base erosion.
ATAD
Profit shifting
Activity shifting
Tax havens
Multinational firms

1. Introduction
In recent years there has been growing concern about aggressive tax
avoidance by multinational corporations (MNCs). In particular, MNCs
are accused of paying virtually no taxes in the places where they operate
by shifting profit to low-tax jurisdictions. These facts have been clearly
documented in the recent literature (e.g., Desai et al., 2006a, Desai et
al., 2006b, Huizinga and Laeven, 2008 and Heckemeyer and Overesch,
2013). Accordingly, in 2013, the OECD launched an initiative against
base erosion and profit shifting (BEPS - OECD, 2013a; b, 2014, 2015).
This initiative, consisting of a 15-point action plan, is intended to facil-
itate intergovernmental cooperation regarding the taxation of multina-
tional corporations. More specifically, the aim is to require the taxing
of profits where the economic activities and value creation occur. In
response to the BEPS, the European Commission (EC) adopted the Anti
Tax Avoidance Package in 2015, which includes a proposal for an Anti
Tax Avoidance Directive (ATAD) (see European Commission, 2016a).1
This directive was adopted on 20 June 2016, and all member states
have to apply these measures from 1 January 2019.
In reality, however, pure profit shifting is not the only way compa-
nies can lower their tax liabilities. They can benefit from low taxation
by shifting real activities abroad if the territorial taxation system,2 cur-
rently used by the majority of OECD countries, applies.3 Importantly,
the BEPS addresses the issue of intangible property (IP) given its poten-
tial risk to contribute to aggressive profit shifting.4 The ATAD also pro-
vides tax exit rules to prevent companies from avoiding tax when relo-

∗ Corresponding author.
E-mail addresses: patrice.pieretti@uni.lu (P. Pieretti), gpulina@gmail.com (G. Pulina).
1
We thank the editor, two anonymous referees, Skerdilajda Zanaj, Luisito Bertinelli and Andreas Haufler for their helpful comments and suggestions.
1
In addition to the ATAD, the EC package consists of three other documents: a Communication for an External Strategy for Effective Taxation, an amendment to
the directive on mutual assistance for the automatic exchange of information in country-by-country reporting, and the recommendation on tax treaties adding the
“genuine economic activity”test to the Principal Purpose Test (PPT) rule (see European Commission, 2016b,c,d).
2
According to this regime, multinationals are not taxed at home on their foreign income.
3
In 2006, the European Court of Justice (ECJ) ruled in the Cadbury Schweppes Case (Case C-196/04)) that the U.K. CFC (controlled foreign company) legislation,
under which the profits of a foreign subsidiary of a resident parent-company are attributed to the resident company and taxed domestically can only apply to wholly
artificial tax minimization schemes. Furthermore, trying to attract foreign businesses with low taxation is consistent with the EU Freedom of Establishment.
4
Dischinger and Riedel (2011) show that moving the location of IP to a low-tax jurisdiction is a simple way used by MNCs to reduce tax liabilities.

https://doi.org/10.1016/j.econmod.2020.01.020
Received 27 March 2019; Received in revised form 22 November 2019; Accepted 25 January 2020
Available online XXX
0264-9993/© 2020 Elsevier B.V. All rights reserved.

Please cite this article as: Pieretti, P., Pulina, G., Does eliminating international profit shifting increase tax revenue in high-tax countries?,
Economic Modelling, https://doi.org/10.1016/j.econmod.2020.01.020
P. Pieretti, G. Pulina
cating assets (e.g., valuable intangible assets) to another jurisdiction.
A post-BEPS/ATAD world will focus on economic substance rather
than the legal ownership of intangibles. Accordingly, a subsidiary will
be entitled to returns from intangibles with respect to economic sub-
stance criteria rather than legal ownership. Therefore, returns from the
exploitation of intangibles will accrue to subsidiaries if there is evidence
of value-creating DEMPE functions.5
In other words, the focus is on artificial profit shifting for tax pur-
poses and not on tax competition per se. In a recent interview in the
Sydney Morning Herald (November 14, 2014), the head of OECD tax
policy told Fairfax Media that the BEPS would “put an end to harm-
ful tax competition, but not to tax competition”. Moreover, Alms et al.
(2016) note that the “… relocation of IP to Ireland can still generate
significant tax savings for U.S. pharmaceutical companies, even in our
post-BEPS world”. While this additional substance is a greater require-
ment than the requirements in the pre-BEPS context, relocating DEMPE
functions and IP ownership allows companies to truly maximize tax sav-
ings while remaining compliant with both the IRS and the local OECD
territory BEPS perspective.
However, low taxes are not always enough to attract foreign firms.
According to Dharmapala (2008) and Dharmapala and Hines (2009),
havens tend to have relatively sophisticated infrastructure6 and Gon-
zalez and Schipke (2011, p.45) observe that “strong regulations that
inspire confidence are a crucial factor” to attract foreign businesses. In
a strict sense, infrastructure refers, for example, to telecommunications
and transportation infrastructure, universities and public R&D opera-
tions. In a larger sense, we must also consider institutional infrastruc-
ture, such as property rights enforcement, capital market regulations
and more generally, attractive laws and regulations. In fact, a country
that invests in the quality of its institutions can be perceived as a safe
location where business activities can be carried out. It is important to
note that many infrastructure services (especially institutional infras-
tructure) have the nature of local public goods (non-rivalry and nonex-
clusion), which benefit all firms in the same location, whether they are
targeted by attractiveness policies, as are MNCs, or not. It follows that
physical and institutional infrastructure together with low taxation are
essential to host foreign activities.
In a recent article, The Economist explains that “to avoid the per-
ception of tax-dodging, some firms are starting to shift executives and
head-office functions to their new domiciles” and that “America is start-
ing to lose not just tax revenues but jobs too”. Consequently, to lower
their tax burdens and avoid the perception of pure profit shifting, MNCs
begin to shift part of their real activity to low-tax jurisdictions (The
Economist, 2015).
While activity shifting does not violate the ongoing anti tax planning
regulation, it still remains a form of base erosion from the point of view
of high-tax countries, as noted by Hines (2014).
The following questions then arise. Do international anti-profit shift-
ing regulations encourage tax havens to intensify their attractiveness to
foreign substance based investments? In this case, will tax havens be
capable of providing more attractive infrastructure? Could this induce
MNCs to engage more in international activity shifting? Is it then con-
ceivable that anti tax avoidance regulations could fail?
This paper develops a simple model to analyze under what condi-
tions initiatives intended to eliminate profit shifting can be successful
given that these actions may induce the relocation of substance-based
activities. In our model, we assume that a representative multinational
company is located in a high-tax region (onshore region) and tries to
reduce its tax liabilities by shifting profit and/or real business to a low-
tax jurisdiction (offshore country). The offshore jurisdiction tries to lure
5
Development, Enhancement, Maintenance, Protection, and Exploitation.
6
Infrastructure can take various forms, such as telecommunications and
transportation infrastructure, universities and public R&D investment, and gov-
ernance infrastructure, such as property rights enforcement, capital market reg-
ulations and more generally attractive laws and regulations.
2
Economic Modelling xxx (xxxx) xxx
multinationals by applying a zero tax rate but requires a fee for profit-
shifting services. Moreover, the offshore jurisdiction invests in attrac-
tive infrastructure to facilitate the setting up of real businesses, under
payment of a fee. The multinational is assumed to maximize its own
net profit by selecting the best mix of the two tax-reducing devices.
The model demonstrates that the removal of tax-motivated profit shift-
ing increases tax revenue in the onshore region if the low-tax juris-
diction is not too efficient in providing attractive infrastructure. This
outcome is more easily achieved when the high-tax country is able to
counter international activity shifting by competing in infrastructure
with the tax haven rather than being passive. Anti tax avoidance regu-
lations have been designed independently of the possible consequences
on tax and nontax competition. It follows that these measures may be
undermined by the emergence of another type of base erosion resulting
from the intensified use of nontax instruments, such as public infrastruc-
ture provision by low-tax countries. Consequently, international regu-
lations aimed at combating aggressive tax avoidance should anticipate
the adverse effects induced by the resulting emergence of other forms
of base erosion.
Moreover, our model is consistent with the existence of two types
of tax havens. One type of haven relies mainly on profit-shifting ser-
vices (e.g., Panama and the Cayman Islands), and the other offers a
mix of profit-shifting and activity-shifting services (e.g., Ireland). We
demonstrate that the cost of shifting activity for tax purposes relative
to shifting profit to the tax haven is crucial in explaining the choice
of the representative MNC between the two types of tax havens. This
result seems to be consistent with reality if we proxy capital mobility
with the relative cost of activity shifting.
Finally, we demonstrate that base erosion and onshore tax revenue
are not necessarily affected in the same way by the elimination of profit
shifting. More precisely, the eradication of profit shifting can increase
tax revenue while not reducing base erosion.
1.1. Contribution and related literature
The paper contributes to the existing literature in three ways.
First, contrary to the previous literature on tax havens, which only
focuses on tax planning and profit-shifting aspects (for example, Slem-
rod and Wilson, 2009; Bucovetsky, 2014; Peralta et al., 2006), our paper
considers that low-tax countries can also be attractive for real business
activities because of the tax advantages they provide. This is in line with
Hines (2014), who observes that “the most straightforward way to be
able to report earning taxable income in low-tax jurisdictions is to con-
centrate economic activity there”. Moreover, he states that “studies con-
sistently find that multinational firms locate more employment, prop-
erty, plants, and equipment in low-tax locations than the structures of
these economies would ordinarily warrant”.
Second, we assume that low-tax jurisdictions respond strategically to
the tax and infrastructure policies of high-tax jurisdictions by choosing
a mix of fees and infrastructure investment. Considering that jurisdic-
tions compete strategically and independently in tax and nontax instru-
ments is consistent with the empirical (Hauptmeier et al., 2012) and
theoretical literature (Justman et al., 2002, 2005; Bénassy-Quéré et al.,
2007; Dembour and Wauthy, 2009; Pieretti and Zanaj, 2011). In this
context, we show that the onshore country is able to mitigate the poten-
tial loss of tax revenue when it strategically adapts its infrastructure
investment (Section 5).
Finally, our paper examines whether policies aimed at the elimi-
nation of aggressive tax avoidance are able to increase tax revenue in
onshore countries, which is ultimately the aim of such initiatives. The
literature on tax havens has, however, concentrated more on policies
intended to eliminate tax havens’ aggressive operations, while implic-
itly assuming that this would coincide with the elimination of tax rev-
enue erosion. An example is Elsayyad and Konrad (2012), who ana-
lyze how the timing of deactivation of the aggressive activity of tax
havens impacts the welfare of the onshore region. In particular, they
P. Pieretti, G. Pulina
conclude that eliminating just the activity of a subset of tax havens may
reduce welfare in the OECD. However, eliminating this activity alto-
gether would increase welfare. Consequently, fully eliminating aggres-
sive tax avoidance is equivalent to an increase in welfare.
Our paper is thus related to the literature that analyzes the role of
tax havens regarding profit shifting and tax competition, as well as the
resulting effect on overall welfare.7 One important result of the pre-
vious literature is that profit shifting lowers the welfare in non-haven
countries (Slemrod and Wilson, 2009). In this context, tax havens do
not compete for investments but only facilitate aggressive tax plan-
ning. However, there are counterarguments according to which non-
havens can benefit from some profit shifting. Johannesen (2010) shows
that profit shifting can eliminate tax rate differences among competing
non-haven countries that differ in size. Hong and Smart (2010) suggest
that international tax avoidance schemes may reduce the tax burdens
of mobile capital thus facilitating investments.
Other prominent contributions argue that profit shifting allows
non-havens to discriminate between mobile and immobile capital (see
Keen, 2001; Peralta et al., 2006; Janeba and Smart, 2006; Bucovet-
sky and Haufler, 2008). According to Hines (2010) this allows high-
tax countries “to maintain high capital tax rates without suffering dra-
matic reductions in foreign direct investment.” Moreover, profit shifting
appears as a substitute for preferential tax treatments, which have come
increasingly under scrutiny.
Generally tax havens are described as providers of tax concealment
services but not as locations of real local activity, tangible or intangi-
ble.8
We depart from the previous literature in different ways. Contrary
to many prominent contributions (e.g., Bucovetsky and Haufler, 2008;
Slemrod and Wilson, 2009; Johannesen, 2010) and similar to Bucovet-
sky (2014), we treat havens as strategic players that are able to design
optimal policies. In addition, we consider that MNCs can benefit from
low taxation by also shifting real activity to tax havens. In other words,
the model we develop is less restrictive since it allows havens to attract
businesses of substance. Finally, we analyze the implications of activ-
ity shifting in the context of the BEPS and the more recent European
response, the ATAD. In fact, these initiatives aim at eliminating profit
shifting but do not consider other sources of base erosion.
The basic purpose of the paper is to analyze under what conditions
anti profit-shifting regulations can make high-tax jurisdictions better
off regarding tax revenue, if we reasonably assume that MNCs can
shift both real activity and profits to a low-tax country. In this con-
text, attractive infrastructure provision as well as tax advantages plays
a critical role.
The next section presents the basic model. In section 3, we assume
that profit shifting can be eliminated. Section 4 compares tax revenue
before and after the removal of profit shifting. Section 5 extends the
model to the case of infrastructure competition between high- and low-
tax jurisdictions. Section 6 concludes the paper.
2. The model set-up
We consider a world economy with two countries, a large high-tax
country H and a small tax haven F.
There are m (normalized to 1) multinational companies (MNCs)
located in country H that produce a homogeneous good that is sold
7 For surveys on the theory of tax havens see Dharmapala (2008) and Keen
and Konrad (2013).
8
The OECD (2015) defines an asset as “intangible” if it is not a physi-
cal/financial asset that can be owned or controlled for use in commercial activ-
ities, and whose use or transfer would be compensated had it occurred in a
transaction between independent parties.
3
Economic Modelling xxx (xxxx) xxx
on a world market to a given price.9 Each MNC has one unit of cap-
ital that generates a profit ΠH = q + 𝛿 . The fraction q results from
firm specific productivity, whereas 𝛿 is the output fraction that depends
on the amount of infrastructure provided in the home country.10 For
convenience and without loss of generality, we set q = 1 so that
ΠH = 1 + 𝛿 .
Assume that the representative MNC contemplates the possibility of
shifting part of its capital to the tax haven in order to set up a substance-
based activity and/or the possibility of offshoring, through a special
purpose entity (SPE), part of its profit generated in H. In reality there
are many ways to shift income from high-tax to low-tax countries. In
order to justify the way we model profit shifting, we assume that capital
is initially shifted to a SPE and then invested back by intra-corporate
loans to the home country, where actual production takes place.11 If
the home country allows firms to deduct interest payments on loans
as an expense item, it follows that interest payments of the MNC to
the SPE located in the tax haven is a means of tax avoidance through
profit shifting. This scheme allows the allocation of income between
affiliates of the same legal entity, regardless where the income has been
generated.
Let 𝜃 denote the fraction of capital moved12 to the tax haven F for
setting up a substance-based activity and 𝜃 p the portion of the profit ΠH
that is merely shifted offshore.
Following Johannesen (2010) and Haufler and Schjelderup (2000),
we consider that profit and capital shifting entail moving costs. We
assume that the cost of activity shifting increases with the amount of
offshored capital,13 while we suppose that the cost of profit shifting
increases with the parameter 𝜃 p that can be viewed as the share of con-
cealed tax-base. The latter assumption that is similar to Haufler and
Schjelderup (2000) results from the fact that additional efforts are nec-
essary to increase income concealment and that the probability of being
detected increases with the degree of concealment.
Setting up a real business activity in a foreign location requires
shifting capital and possibly specific human resources. We introduce
a convex cost function to account for the costs involved in the relo-
cation of productive resources to a foreign country. Among these, we
have to consider moving costs, such as freight costs, storing costs and
insurance fees. Moreover, there are costs of renting or buying land and
business premises, and costs of relocating high skilled employees (who
are hardly available in the host location) in a foreign country. Mov-
ing costs may increase proportionally with the amount of shifted cap-
ital (business). However, the cost of renting land and premises will
increase more than proportionally with the magnitude of business if
extra physical space is scarce. Finally, given that additional activity
requires more skilled people, it is plausible that relocating employees
becomes increasingly difficult when more business is shifted abroad (for
example, difficulties to find people who adapt to a new country with a
different language and a different culture).
9 We assume that there are a large number of MNCs located in country H.
Consequently, this rules out individual bargaining situations between the MNCs
and the jurisdictions.
10
Note that our profit function is a reduced form that does not exclude the
existence of a production function. Also, note that the total amount of capi-
tal that can be invested either at home or abroad is given and normalized to
1. However, this normalization does not prevent that the amount of capital
invested at home may be an endogenous variable.
11
According to UNCTAD (2015), a big part of FDI flowing into tax havens are
not invested in projects based in the country. Instead, they can flow back to the
country of source, which is a process called “round tripping”. As an example
UNCTAD (2013) reports that “[..] the top three destinations of FDI flows from
the Russian Federation - Cyprus, the Netherlands and the British Virgin Islands
- coincide with the top three investors in the Russian Federation [.. ] “.
12
It follows that the profit generated at home equals (1 −𝜃 )𝛱 H .
13
Offshoring real activities involves costs resulting from building or renting
the business premises as well as from moving the starting material, etc.
P. Pieretti, G. Pulina
We introduce convex costs for profit and activity shifting. The cor-
responding cost functions are respectively 12 𝜃p2 and 𝛼2 𝜃 2 . Note that it is
common practice to use convex functions in two-country models (Hau-
fler and Schjelderup, 2000 and Stöwhase, 2005) that try to capture the
costly effort of MNCs to avoid home taxation. The parameter 𝛼 ∈ R+
accounts for the difficulty to shift activity relative to profit. Because
profit shifting can be achieved virtually, we can assume that shifting
activities is at least as difficult. In other terms, we assume that 𝛼 > 1.14
The onshore country taxes profit at a proportional rate t ∈ [0, 1].
Generally, it is not very costly to offer tax sheltering services, but pub-
lic investments are required to guarantee a favorable environment to
attract foreign firms that invest in substance-based activities. In this
context, we do not only consider public investments in physical assets,
but also institutional infrastructure such as property rights enforce-
ment, capital market regulations, labor and environmental regulations,
accounting standards and disclosure requirements.
Consequently, let G be the amount of public infrastructure provided
by the tax haven. Let us assume that G units of infrastructure are needed
to generate G units of net output. Additionally, we suppose that the
cost of infrastructure provision increases at an increasing rate due to
diminishing returns to scale. We thus introduce a quadratic function
𝛽 2
2
G , where 𝛽 measures the efficiency of the tax haven in providing
attractive infrastructure. More exactly, the coefficient 𝛽 accounts for the
capacity of a jurisdiction to cope with key stages of public infrastructure
investments, which consist of planning, allocation and implementation.
The higher the value of 𝛽 , the lower the efficiency in providing G. Given
that the size of 𝛽 plays a key role in the main conclusions of the paper,
the following clarifications are needed.
For Pritchett (2000), “the difference between investment cost and
capital value is of first-order empirical importance”. In this regard, a
growing body of literature emphasizes the importance of legal, insti-
tutional, and procedural arrangements for public investment manage-
ment. Recent assessments made by the IMF (2015) highlight average
inefficiencies in public investment processes of around 30 percent.15
In the same vein, Esfahani and Ramirez (2003) show, by estimating
a structural model of infrastructure and growth, that achieving bet-
ter outcomes requires specific institutional and organizational condi-
tions, which are “more fundamental than simply designing infrastruc-
ture projects and spending money on them”. For example, Kenny (2009)
reports differences in the building cost of a kilometer of similar roads
that vary by five to ten times across different countries.16
The tax haven levies fees for the tax planning and infrastructure ser-
vices it provides. Accordingly, a fee fp is charged on one unit of capital
accruing to the SPE to shift profit abroad, and a fee f is levied on one
unit of capital invested in the tax haven. Tax havens may charge dif-
ferent types of fees. For instance, Cayman Islands17 charge an annual
fee that is proportional to affiliates’ capital. Similarly, affiliates estab-
lished in Gibraltar pay a set-up fee of 0.5% on authorized capital. In
Luxembourg, funds have to pay an annual subscription tax at the rate
of 0.05% on their total assets.
By analogy to what we did above, capital shifted to the tax haven
yields a net output depending on both firm and jurisdictional charac-
teristics. It follows that a fraction 𝜃 of capital moved to the tax haven
generates a profit equal to 𝜃ΠF , where ΠF = 1 + G.
14
If 0 < 𝛼 < 1, activity shifting would be relatively easier to achieve. This
would make the removal of profit shifting less able to augment tax revenues in
high-tax countries.
15
Note that such inefficiencies are not particular to developing countries.
16
Regarding our cost function 𝛽2 G2 , this means that for a given value of G the
coefficient 𝛽 can be vary across different countries.
17
Cayman Islands levies no taxes on income.
4
Economic Modelling xxx (xxxx) xxx
We can now write the following profit equation of the representative
MNC
1 𝛼
V = (1 − 𝜃p − 𝜃)ΠH (1 − t ) + 𝜃p (ΠH − fp ) + 𝜃(ΠF − f ) − 𝜃p2 − 𝜃 2 . (1)
2 2
The first term of (1) represents the net (of corporate tax) income
generated at home (country H). The second and third terms refer to the
net income resulting respectively from profit and activity shifting.18
Finally, the two last terms are respectively the cost of offshoring profit
and activity. Note that MNC gross-profit is endogenously determined
by the capital allocation decision of the MNC and the infrastructure
investments of the tax haven and the onshore country. Moreover, in
our setting, profit shifting has a positive effect on the capital invested
onshore. In fact, the easier shifting profit offshore becomes, the lower
will be the incentive to shift activity abroad. As a consequence, the
incentive to invest capital onshore will be higher. This positive effect
has also been highlighted by previous studies such as Overesch (2009)
and Desai et al. (2009).
Let us now consider a two-stage game. We assume that the juris-
dictions set their policy instruments in a first stage and that the MNC
decides on its location in a second stage. In fact, it is common that MNCs
decide to locate in a given country after having been informed, in par-
ticular, about the tax rules that the national tax authorities of the host
country apply. The national tax authorities guarantee the enforcement
of these rules in binding agreements (advance pricing agreements),
which allow the MNCs to know the tax conditions they are going to face
before their investment decision. Therefore, country H sets the corpo-
rate tax rate t while the tax haven F sets the fees fp and f and the infras-
tructure provision G. These last choices are made simultaneously and
non-cooperatively by the jurisdictions. Finally, the MNC decides on 𝜃 p
and 𝜃 , namely, the shares of profit and activity shifting respectively.19
2.1. Reallocation of taxable income and activity shifting
In this section, we analyze how a representative MNC located in
country H decides on the mix of taxable income allocation and activity
shifting. Consequently, after having done some substitutions in (1), we
assume that the MNC maximizes V(𝜃 p , 𝜃) with respect to 𝜃 p and 𝜃 :
max V (𝜃p , 𝜃) = (1 − 𝜃p − 𝜃)(1 + 𝛿)(1 − t ) + 𝜃p (1 + 𝛿 − fp )
𝜃p ,𝜃
1 2 𝛼 2
+ 𝜃(1 + G − f ) − 𝜃 − 𝜃 , (2)
2 p 2
subject to
𝜃p + 𝜃 ≤ 1 , (1a)
𝜃p ≥ 0, (1b)
𝜃 ≥ 0. (1c)
18
In the main text we assume that the fee fp is charged for a unit of capital
shifted to the SPE. However, even if we assume fp charged on shifted profit, the
model solves almost identically.
19
In the first part of our model, the number of policy instruments designed
to improve a country’s economic attractiveness is higher for the tax haven rel-
atively to the onshore country. This assumption reflects the fact that havens
are often small countries that lack productive domestic resources. It follows
that these countries use policy variables in order to lure foreign capital an pro-
vide tax mitigation services. Very large countries are less dependent on external
resources. This is reflected in this paper by assuming an uneven number of pol-
icy instruments used to attract/retain firms. We relax this assumption in Section
5, which assumes that the onshore region adopts a more reactive attitude by
also using infrastructure as a possible way to retain mobile firms. However, we
do not exclude that large countries implement other policy variables that are
not relevant to the purpose of our paper.
P. Pieretti, G. Pulina Economic Modelling xxx (xxxx) xxx

The first order conditions resulting from the above maximization shifting by MNCs. Accordingly, the tax haven has an incentive to cap-
yield ture a part of firms’ resulting tax gains by increasing its fees.
Solving the system of the above best responses, we obtain the fol-
𝜃p = t (1 + 𝛿) − fp , (3)
lowing equilibrium values
2𝛼(2𝛼𝛽 − 1 + 𝛽𝛿)
t∗ = ,
t (1 + 𝛿) − f + G − 𝛿 (2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1))(1 + 𝛿)
𝜃= . (4)
𝛼 𝛼(2𝛼𝛽 − 1 + 𝛽𝛿)
fp∗ = ,
Equation (4) highlights that a multinational company can have an 2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1)
incentive to invest in real offshore activities even if the home coun- 𝛼𝛽(2𝛼 − 𝛿(2 + 3𝛼))
try provides more attractive infrastructure relative to the tax haven f∗ = ,
2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1)
(𝛿 > G). A sufficient condition for the occurrence of activity shifting is
that the tax gain t(1 + 𝛿) − f has to be higher than 𝛿 − G. 2𝛼 − 𝛿(2 + 3𝛼)
G∗ = .
2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1)
2.2. The governments’ decisions An interior solution21 is guaranteed by the conditions
1 2𝛼
We assume that the onshore country sets the corporate tax rate t 𝛼𝛽 > and 𝛿< . (5)
2 2 + 3𝛼
that maximizes its tax revenue T. This assumption is similar to that
The equilibrium shares of profit and activity shifting are
used by Kanbur and Keen (1993), Zissimos and Wooders (2008) and
Pieretti and Zanaj (2011). In fact, if one considers that jurisdictions 𝛽(2𝛼 − 𝛿(2 + 3𝛼))
𝜃∗ = ,
are not self-interested governments, it is reasonable to assume that the 2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1)
collected taxes are used to finance public consumption. This objective 𝛼(2𝛼𝛽 − 1 + 𝛽𝛿)
is consistent with the classical welfarist view in which consumers place 𝜃p∗ = .
2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1)
a very high marginal valuation on public goods, which are financed by
tax revenue (see Kanbur and Keen, 1993). The equilibrium tax revenue in the high-tax country is
As it is often the case, we assume that the tax haven is a small 4𝛼(1 + 𝛼)(2𝛼𝛽 − 1 + 𝛽𝛿)2
T∗ = . (6)
country whose source of revenue is only derived from foreign affiliates. (2(3𝛼𝛽 − 1) + 3𝛼(2𝛼𝛽 − 1))2
Therefore, the tax haven decides on the fees fp and f and the infrastruc-
ture expenditure G that maximize its net revenue W. The optimization
problems are 3. The elimination of profit shifting

max T = t (1 − 𝜃p − 𝜃)Πh , In this section, we analyze how the elimination of profit shifting
t
can affect tax revenues in the high-tax country H. In other words, we
𝛽
max W = fp 𝜃p + f 𝜃 − G2 . assume that profit shifting is not anymore an option for the MNC and
fp ,f ,G 2
therefore we set 𝜃 p = 0.
subject to the non negativity conditions (1a)-(1c). As the share of capital invested offshore is now the only available
While maximizing their respective objectives, the governments strategy, the MNC’s profit function becomes
anticipate the MNC’s behavior. Taking account of (3) and (4), the best ̂ = (1 − 𝜃)(1 + 𝛿)(1 − t ) + 𝜃(1 + G − f ) − 𝛼
V 𝜃2.
response of the onshore jurisdiction becomes 2
𝛼(1 + fp ) + f − (G − 𝛿) Maximizing the profit function relative to 𝜃 for given t, f and G yields
t br = . the share of activity shifting
2(1 + 𝛼)(1 + 𝛿)
t (1 + 𝛿) − f + G − 𝛿
The onshore country’s best tax response t is increasing in the tax haven’s 𝜃̂ = .
fees and decreasing in the haven’s infrastructure provision G. Intu- 𝛼
itively, the onshore country has an incentive to increase (decrease) its The governments’ best responses are
tax rate any time the tax haven decreases (increases) its attractiveness.
Accordingly, the best responses of the tax haven are ̂t br = 𝛼 + f − (G − 𝛿) ,
2(1 + 𝛿)
t
fpbr = (1 + 𝛿), ̂f br = 𝛼𝛽(t (1 + 𝛿) − 𝛿) ,
2
2𝛼𝛽 − 1
𝛼𝛽(t (1 + 𝛿) − 𝛿)
f br = , t (1 + 𝛿) − 𝛿
2𝛼𝛽 − 1 ̂ br =
G .
2𝛼𝛽 − 1
t (1 + 𝛿) − 𝛿
Gbr = . Solving the above equation system yields the following equilibrium
2𝛼𝛽 − 1
variables
The second order conditions are always satisfied for 𝛼𝛽 > 12 .20 In other
words, we assume that activity shifting and providing attractive infras- ̂t ∗ = 𝛼(2𝛼𝛽 − 1 + 𝛽𝛿) , (7)
(3𝛼𝛽 − 1)(1 + 𝛿)
tructure are jointly difficult enough. Note that when the onshore econ-
omy increases its corporate tax rate, the tax haven’s best response is to ̂f ∗ = 𝛼𝛽(𝛼 − 𝛿) , (8)
raise its infrastructure provision and the fees of the services it provides. 3𝛼𝛽 − 1
The reason is because higher taxation encourages profit and activity
̂∗ = 𝛼−𝛿
G . (9)
3𝛼𝛽 − 1

𝜕2 T 𝜕2 W 𝜕2 W 𝜕2 W
20
It is easy to check that 𝜕t 2
< 0 for all t ∈ [0, 1] and 𝜕f 2
< 0, 𝜕 fp2
< 0, 𝜕 G2
<
𝜕2 W 𝜕2 W 𝜕2 W 2
0. Moreover, 𝜕f 2
· 𝜕 G2
− ( 𝜕f 𝜕G
) >0 if 𝛼𝛽 > 1
2
. 21
It follows that condition 1a is not binding.

5
P. Pieretti, G. Pulina Economic Modelling xxx (xxxx) xxx

These equilibrium variables are positive because of condition (5). nation of profit shifting increases activity shifting relatively more than
The share of activity shifting becomes corporate taxation, tax revenue will decrease. Consequently, the elim-
ination of profit shifting translates into higher onshore tax revenues
𝛽(𝛼 − 𝛿)
𝜃̂∗ = . (10) if tax havens are unable to increase sufficiently their attractiveness to
3𝛼𝛽 − 1
foreign real business.
Note that the elimination of profit shifting increases the share of real The previous proposition shows that the total elimination of profit
business shifted. Indeed, it is easy to show that 𝜃̂∗ − 𝜃 ∗ > 0 given that shifting, which is ultimately destined to increase tax revenue in high-tax
𝛼𝛽 > 12 .22 countries, may not be successful given that it triggers another form of
Finally, the equilibrium tax revenue in the high-tax country equals base erosion, namely activity shifting. However, profit shifting may not
𝛼(2𝛼𝛽 − 1 + 𝛽𝛿)2 be entirely eliminated. In fact, current actions and regulations, such as
̂∗ =
T . (11)
(3𝛼𝛽 − 1)2 the BEPS and the ATAD, allow some profit shifting to low-tax jurisdic-
tions. Moreover, shifted profit can be subject to a minimal taxation in
low-tax jurisdictions (see the “Controlled Foreign Company (CFC) rule”
4. The effectiveness of removing profit shifting in the ATAD Article 7(1)).
Our model can be easily adapted to include a binding cap on profit
Now, we compare the results derived from the non-profit shifting shifting and the imposition of a minimum offshore tax. Then, it can be
scenario of the previous section with those of the bench-mark scenario shown that the main results of the paper, highlighted in Proposition 1,
of Section 2. In the following proposition we establish under what con- still remain.24
dition removing profit shifting can have an adverse effect on the tax Finally, the following proposition highlights cases in which the MNC
revenue in the high-tax country. prefers it when competition is restricted to being only in productive
Proposition 1. The removal of profit shifting decreases the tax revenue in investments.
the high-tax country if the haven is efficient enough in providing attractive Proposition 2. When the tax haven is efficient enough in providing infras-
infrastructure . However, this is less likely to occur the more it is costly to tructure and the provision of infrastructure in the onshore country is not too
shift activities in the tax haven. Formally: ΔT = T̂∗ − T ∗ < 0 if 𝛽 < 𝛽̃, high, the MNC is better off when profit shifting is eliminated. Formally,
where
( )
∗ >0
Va∗ − Vpa ̂
for 𝛽 < 𝛽.
̃ 1 1 1 𝜕 𝛽̃
𝛽= + √ and <0. Proof. In Appendix B.
𝛼 2 6 1+𝛼 𝜕𝛼 ▪

The intuition is that the elimination of profit shifting reduces the

Proof. Let us set
intensity of tax competition for mobile firms and increases infrastruc-
ΔT = A · B, ture competition. The MNC can gain, swapping profit shifting for activ-
ity shifting, as long as the tax haven is very efficient and infrastructure
where provision is not too high in the home country.
𝛼 2 (2𝛼𝛽 − 1 + 𝛽𝛿)2
A= > 0, (12)
(3𝛼𝛽 − 1)2 (6𝛼𝛽(1 + 𝛼) − (2 + 3𝛼))2 5. Infrastructure competition

In this section, we assume that the high-tax region tries to counter

B = 36(1 + 𝛼)𝛼 2 𝛽 2 − 36(1 + 𝛼)𝛼𝛽 + 8 + 9𝛼.
It follows that sign (ΔT ) = sign (B). From (13), we see that B is quadratic
in 𝛽 . Solving B(𝛽) = 0 relative to 𝛽 yields two distinct positive roots,23
but only the highest value 𝛽̃ of both is relevant if we impose condition
(5) that guarantees the existence of an interior solution. So, we conclude
that
{
≤ 0 𝛽 ≤ 𝛽, ̃
ΔT =
>0 otherwise.▪
Intuitively, the removal of profit shifting makes MNCs more captive,
which induces the onshore region to increase its corporate tax rate t.
This entices the tax haven to invest in attractive infrastructure designed
to encourage activity shifting by MNCs which increases 𝜃 . If the elimi-
(13) international activity shifting by competing in infrastructure with the
tax haven. One important aspect is to check the robustness of the above
conclusions by endogenizing the infrastructure choice of the onshore
country.
In the following, we limit ourselves to the main results and the
computations are relegated to Appendix A. For consistency reasons, we
redefine 𝛿 as Gh and G as Gf when there is infrastructure competition.
As we did it for the tax haven, we assume that Gh units of infrastruc-
ture are needed to generate Gh units of net output in the onshore econ-
omy. Hence, Gh accounts for both the infrastructure provision and the
resulting output that accrues to home investing firms. Once again, we
assume that the cost of infrastructure provision is given by an increasing
G 2
convex function 2h .
Now the onshore country is supposed to choose the corporate tax

22
More generally, it is possible to demonstrate that in equilibrium the share
of activity and profit shifting are negatively related. To show this, assume that 24
A more general model (available on request) leads to the following con-
the share of profit shifting is bounded from above. The “capped” share of profit clusions. A positive cap on profit shifting rather than the complete elimination
shifting becomes 𝜃p (with 𝜃p ≤ 𝜃p < 𝜃p∗ ). A decrease in 𝜃p can be interpreted as of profit shifting (0 < 𝜃p ≤ 𝜃p ) decreases tax revenue in the onshore country.
a more restrictive policy regarding profit shifting devices, while an increase in Moreover, requiring a minimum offshore tax on shifted profits will result in
𝜃p makes profit shifting easier. After having solved the model with this binding two opposing effects. First, it will reduce profit shifting (positive effect) and
𝜕 𝜃̂∗
cap it is possible to demonstrate that
𝜕𝜃p
< 0 given that 𝛼𝛽 > 12 . It follows that, second, it will induce additional activity shifting (negative effect). Which effect
ultimately dominates depends on the parameter constellation of the model and
the easier shifting profit offshore becomes (increasing 𝜃p ), the lower will be the
in particular, on the efficiency of the tax haven in providing attractive infras-
incentive to shift activity abroad (lower 𝜃̂∗ ). tructure. Finally, note that the more constraining these two regulations, the
( ) ( )
23 ̃
𝛽 = 𝛼1 12 + √1 > 0 and ̃𝛽̃ = 𝛼1 12 − √
1
> 0. more they will be equivalent to the total elimination of profit shifting.
6 1+𝛼 6 1+𝛼

6
P. Pieretti, G. Pulina
rate t and the amount Gh . Consequently, the problem of the high-tax
country becomes25
Gh 2
max Wh = t (1 − 𝜃p − 𝜃)(1 + Gh ) − . of immobile firms is relatively higher.
t ,Gh 2
As already assumed in the previous sections, the tax haven is sup-
posed to decide on the fees fp and f and the infrastructure expendi-
ture Gf . In the Appendix we solve the game between the onshore and
offshore jurisdictions with and without profit shifting. The following
results can then be highlighted.26
First, we show (see Appendix A) that the onshore economy can
provide a higher equilibrium amount of infrastructure than the tax
haven (Gh ∗∗ > Gf ∗∗ and G ̂h ∗∗ > G
̂f ∗∗ ) if 𝛽 > 1 . Notwithstanding this
2
case, the onshore economy is not able to prevent activity shifting
(𝜃 ∗∗ > 0 and 𝜃̂∗∗ > 0). As we explained in a previous section, this is
the case because the relative tax advantage provided by the tax haven
is high enough.
Second, in case of a removal of profit shifting the welfare of the high-
tax economy will not unambiguously increase. Indeed, the elimination
of tax planning reduces the onshore welfare (the tax revenue net of the
cost of infrastructure provision) if the low-tax jurisdiction is efficient
enough in providing attractive infrastructure (W ̂h ∗∗ < W ∗∗ if 𝛽 < 𝛽̃′ ).
h
However, when the high-tax economy does not compete in infrastruc-
ture provision, the threshold under which the level 𝛽 induces a wel-
fare decrease is higher than in the case of infrastructure competition.
Indeed, we prove in Appendix A that 𝛽̃′ < 𝛽̃.
Proposition 3. When the high-tax jurisdiction uses infrastructure provi-
sion strategically, the removal of profit shifting can reduce the welfare in the
high-tax economy if 𝛽 < 𝛽̃′ , namely when the tax haven is efficient( enough
) 𝛼(2𝛼𝛽 − 1)(1 − 2𝛽)
in providing attractive infrastructure. However, for each 𝛽 ∈ 𝛽̃′ , 𝛽̃ the
elimination of profit shifting is beneficial to the high-tax region if it competes
in infrastructure and detrimental when it does not. Formally,
if 𝛽̃ > 𝛽 > 𝛽̃′ , it follows that W
̂h ∗∗ > W ∗∗ and T
̂∗ < T ∗ .
h increasing welfare in the high-tax jurisdiction because 𝛽 > 𝛽̃′ , given
Proof. In Appendix A.
The underlying intuition is the same as explained above. The elim-
ination of profit shifting has two effects. On the one hand, it increases
the incentive to raise the tax rate in the high-tax region, but on the
other hand, it induces more activity shifting. The second effect domi-
nates when the low-tax country is efficient enough in attracting foreign
businesses.
A more general and realistic approach should account for the exis-
tence of internationally immobile firms in country H in addition to
MNCs. Given that infrastructure facilities are public goods, we could
account for positive spillovers induced by the infrastructure provision
on the productivity of the whole economy. Moreover, the model could
also account for the size of the onshore jurisdiction, if we reasonably
assume that the country size and the share of internationally immobile
firms in the total number of domestic firms are positively correlated.
In this case, the bigger the country H, the less it is likely to condition
its corporate tax rate on the policy responses of MNCs and foreign tax
havens. The corporate tax rate will then be tailored to all the domestic
firms and not only on MNC′s activities.
In Appendix C, we incorporate internationally immobile firms in
our model. It is then possible to demonstrate that Proposition 3 does
not change. However, if the tax haven is efficient enough in provid-
ing attractive infrastructure (𝛽 < 𝛽̃′ ), it can be shown that the higher
the number of immobile firms (or the bigger the country H), the larger
25
It is easy to check that the following second order conditions are verified:
𝜕 2 Wh
< 0, 𝜕𝜕GW2h < 0 and 𝜕𝜕W · 𝜕𝜕GW2h − ( 𝜕𝜕t 𝜕WGh )2 > 0, given that 𝛼 ≥ 1.
2 2 2 2
h
𝜕t 2 h t2 h h
26
We use the suffix ∗∗ to refer to the case of infrastructure competition and
the circumflex ̂ is used in the absence of profit shifting.
Economic Modelling xxx (xxxx) xxx
the loss of tax revenue after the elimination of profit shifting. By con-
trast, if the tax haven is not efficient enough (𝛽 > 𝛽̃′ ), the elimination of
profit shifting results in higher gains in tax revenue when the number
The underlying intuition can be explained as follows. The higher
the number of immobile firms, the less the onshore tax policy will be
sensitive to MNCs’ behavior. This tends to increase the home tax rate.
The impact of this increase on the domestic tax revenue depends on
how efficient the haven is in providing attractive infrastructure. If it
is very efficient (𝛽 < 𝛽̃′ ), the outflow of capital to the tax haven will
dominate the home tax increase. As a result, the tax revenue decreases.
Otherwise (𝛽 > 𝛽̃′ ), the home tax increase will dominate the capital
outflow and consequently, tax revenue increases in the country H with
the elimination of profit shifting.
Now, notice that the BEPS initiative focuses on the concept of base
erosion. In our model, base erosion () is given by the parameters 𝜃 p
and 𝜃 , which stand respectively for the shares of profit and activity
shifting. Consequently, the amount of base erosion with and without
profit shifting (in case of infrastructure competition) is given by
𝛼(2𝛼𝛽 − 1) + 2𝛽(𝛼 − 1)
 = 𝜃 ∗∗ + 𝜃p∗∗ = , (15)
3𝛼(2𝛼𝛽 − 1) + 2d
̂ = 𝜃̂∗∗ = 𝛽(𝛼 − 1)
 , (16)
d
where d = 𝛽 (3𝛼 − 1) − 1.
Let us now check if base erosion and tax revenue losses incurred
by high-tax jurisdictions move in the same direction. To this end we
consider the difference Δ = ̂ −  given by the following expression.
Δ = . (17)
d(3𝛼(2𝛼𝛽 − 1) + 2d)
It follows that the removal of profit shifting induces base repatria-
tion (Δ < 0) if the tax haven is not efficient enough (𝛽 > 12 ) in pro-
viding attractive infrastructure. In this case, base repatriation implies
(14)
that 12 > 𝛽̃′ .27 Note that a decrease in base erosion is sufficient to induce
▪
increased tax revenue but it is not a necessary condition. Indeed, if
𝛽̃′ < 𝛽 < 12 , the eradication of profit shifting increases tax revenue in
the onshore economy even if base erosion increases. The intuition can
be explained as follows.
When profit shifting is eliminated, firms are more exposed to higher
taxation in their home country. As a reaction, the tax haven tries to
encourage activity shifting by increasing its infrastructure provision.
This can result in a higher outflow of taxable income such that base
erosion increases but to a lesser extent than taxation increases as a
result of the removal of profit shifting. This is precisely what happens
when the efficiency parameter is larger than 𝛽̃′ but smaller than 12 .
Proposition 4. A reduction of base erosion as a consequence of the
removal of profit shifting is sufficient to induce increased tax revenue in
the onshore economy. The reverse is not necessarily true. In particular, if
𝛽̃′ < 𝛽 < 12 , the eradication of profit shifting increases tax revenue in the
onshore economy even if base erosion increases.
Proof. In the text.
6. Conclusion
Base erosion through profit shifting is considered a serious problem
since it potentially reduces tax revenue in onshore economies. In this
𝜕 𝛽̃′
27
Indeed, lim𝛼 →1 𝛽̃′ = 1
2
and 𝜕𝛼
≤ 0, ∀𝛼 > 1.
7
P. Pieretti, G. Pulina Economic Modelling xxx (xxxx) xxx

context, the BEPS initiative, launched recently by the OECD (OECD,
2013a; b, 2014), and the European directive, the ATAD, should help
eliminate profit shifting with the aim of increasing onshore tax revenue.
This paper demonstrates that the positive effects of the BEPS initia-
tive and the ATAD could be undermined. In fact, in the same vein as
Hines (2014), we argue that anti tax planning regulations can induce
multinational firms to use other ways to mitigate their tax liabilities,
such as shifting real activities to tax havens. As a result, the removal of
tax-motivated profit shifting increases the tax revenue in the onshore
region only if the low-tax jurisdiction is not too efficient in providing
attractive infrastructure. Additionally, we highlight that the BEPS ini-
tiative and its regional implementations, such as the ATAD, are more
likely to be successful if the high-tax country is able to counter the shift-
ing of international activity by competing in infrastructure rather than
being passive.
Neither the BEPS initiative nor the ATAD account for international
activity transfers that could result from anti profit-shifting campaigns.
However, we could witness the emergence of another form of base ero-
sion that deserves official scrutiny. Of course, tax havens must be able
to attract foreign business activity and therefore must provide sufficient
business-friendly infrastructure. Moreover, according to their character-
istics they can only offer certain types of infrastructure services. Indeed,
given that most tax havens tend to be small in size and thus lack nat-
ural and human resources and have limited land area, they have few
options beyond financial services and tourism (Morriss, 2008). An effi-
cient way to specialize in offshore financial activities can be achieved
through attractive institutional infrastructure provision. The small size
of their governments makes them more responsive to changes in the
international economic environment and consequently, they tend to be
efficient in adapting and innovating legislation and regulations to facili-
tate financial transactions. According to Morriss (2008), “The most suc-
cessful offshore jurisdictions today have evolved from an initial focus
on company registries and similar activities to a mix of financial cen-
ter activities that include legal, accounting, and other services in an
attempt to expand the portion of the economic activity occurring within
their borders”.
Finally, we also highlight that fighting aggressive tax planning does
not necessarily affect base erosion and onshore tax revenue in the same
way. In particular, a possible eradication of profit shifting can increase
tax revenue without the need to reduce base erosion.

Appendix A. the model with infrastructure competition

First we solve the game described in Section 5 in the absence of profit shifting. The resulting equilibrium variables are given as follows:

̂t ∗∗ 𝛼(2𝛼𝛽 − 1)
= ,
d + 2𝛼𝛽 − 1

̂h ∗∗ = 2𝛼𝛽 − 1
G ,
d
̂f ∗∗ = 𝛼−1
G ,
d
̂f ∗∗ = 𝛼𝛽 · G
̂f ∗∗ ,

𝛽(𝛼 − 1)
𝜃̂∗∗ = ,
d
where d = 𝛽(3𝛼 − 1) − 1 > 0 since 𝛼 ≥ 1 and 𝛼𝛽 > 1
2
.
̂h ∗∗ > G
Note that when 𝛽 > 12 , the onshore economy provides more attractive infrastructure than the tax haven (G ̂f ∗∗ ).
The equilibrium tax revenue in the high-tax country is

̂h ∗∗ = (2𝛼 − 1)(2𝛼𝛽 − 1)2

W . (18)
2d2
If we consider the possibility of profit shifting, the game yields the following equilibrium results
( )
2d + 2(2𝛼𝛽 − 1) ̂t ∗∗ < ̂t ∗∗ ,
t ∗∗ =
(3𝛼 + 2)(2𝛼𝛽 − 1) + 2d
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
<1
( )
2d ̂h ∗∗ < G
̂h ∗∗ ,
Gh ∗∗ = G
3𝛼(2𝛼𝛽 − 1) + 2d
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
<1
( )
2d ̂f ∗∗ < G
̂f ∗∗ ,
Gf ∗∗ = G
3𝛼(2𝛼𝛽 − 1) + 2d
⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟
<1

f ∗∗ = 𝛼𝛽 · Gf ∗∗ ,

t ∗∗
fp∗∗ = (1 + Gh ∗∗ ).
2
The equilibrium levels of profit and activity shifting are
2𝛽(𝛼 − 1)
𝜃 ∗∗ = ,
3𝛼(2𝛼𝛽 − 1) + 2d
𝛼(2𝛼𝛽 − 1)
𝜃p∗∗ = .
3𝛼(2𝛼𝛽 − 1) + 2d
The equilibrium tax revenue in the high-tax country is
2(2𝛼(1 + 𝛼) − 1)(2𝛼𝛽 − 1)2
Wh∗∗ = ,
(3𝛼(2𝛼𝛽 − 1) + 2d)2

8
P. Pieretti, G. Pulina Economic Modelling xxx (xxxx) xxx

It follows that
̂h ∗∗ − W ∗∗ < 0 if
ΔWh = W h
𝛽 < 𝛽̃′
To provide a proof, first set ΔWh = A′ · B′ , where

1 (2𝛼𝛽 − 1)2
A′ = [ ] > 0, (19)
2 d2 3𝛼(2𝛼𝛽 − 1) + 2d 2

[ ]
B′ = 4 4 + 9𝛼(−2 + 𝛼 + 2𝛼 2 ) 𝛼 2 𝛽 2 (20)

[ ]
− 4 3 + 𝛼(9𝛼(1 + 2𝛼)) − 17 𝛼𝛽
[ ]
+ 𝛼 𝛼(7 + 18𝛼) − 12 .
( )
It follows that sign (ΔWh ) = sign B′ . From (20), we see that B′ is quadratic in 𝛽 . Solving B′ (𝛽) = 0 relative to 𝛽 yields two distinct positive
roots,28 but only the highest value 𝛽̃′ of both is relevant if we impose the condition (5) that guarantees the existence of an interior solution.
So, we conclude that
{
≤0 𝛽 ≤ 𝛽̃′ ,
ΔWh =
>0 otherwise.

Recalling the threshold value 𝛽̃, it follows that 𝛽̃ > 𝛽̃′ > 0 for finite values of 𝛼 (𝛼 > 1). The proof is derived from the following properties. The
𝜕 𝛽̃ 𝛽̃′
difference 𝛽̃ − 𝛽̃′ is strictly positive when 𝛼 = 1 and tends to 0 when 𝛼 → +∞. In addition, we verify that 𝜕𝛼 < 0 and 𝜕𝜕𝛼 < 0, ∀𝛼 > 1.

Appendix B. proof of Proposition 2

If competition is restricted to being only in productive investments, the representative multinational can only decide on activity shifting. In this
case, the profit is
( )
𝛼𝛽 2 18𝛼 − 11𝛼 2 + 𝛿 2 + 10𝛼𝛿 + 2𝛼𝛽 (5𝛼 − 5𝛿 − 6) − 2 (𝛼 − 1) + 2𝛿
Va∗ = .
2(3𝛼𝛽 − 1)2
If competition is not restricted to being only in productive investments, the representative multinational can decide on activity and profit shifting. In
this case, the profit is

∗ 𝛼(1 + 𝛼)(−4𝛿 2 (11 + 7𝛿) + 4𝛼 2 (7 + (17 − 6𝛿)𝛿) + 𝛼(72 + 𝛿(16 − 𝛿(31 + 39𝛿))))𝛽 2
Vpa = +
2[3𝛼 + 2 − 6𝛼𝛽 (𝛼 + 1)]2 (1 + 𝛿)
(−4𝛼(2 + 𝛼)(6 + 7𝛼) − 2𝛼(4 + 𝛼(39 + 34𝛼))𝛿 + 2(8 + 𝛼(24 + 𝛼(29 + 12𝛼)))𝛿 2 + 8(1 + 𝛼)(2 + 3𝛼)𝛿 3 )𝛽
+
2[3𝛼 + 2 − 6𝛼𝛽 (𝛼 + 1)]2 (1 + 𝛿)
+8 + 𝛼(16 + 7𝛼) + 𝛼(16 + 17𝛼)𝛿 − 2(2 + 𝛼)(2 + 3𝛼)𝛿 2
.
2[3𝛼 + 2 − 6𝛼𝛽 (𝛼 + 1)]2 (1 + 𝛿)

The representative multinational prefers when competition is restricted to being only in productive investments if Va∗ > Vpa ∗ . This is the case when

the tax haven is efficient enough in providing infrastructure and the provision of infrastructure in the onshore country is low enough.
The proof is given as follows. After having computed the difference Va∗ − Vpa ∗ it follows that the sign of V ∗ − V ∗ equals the sign of the following
a pa
polynomial:
( )
P(𝛽) = 9𝛼 3 (1 + 𝛼)(2𝛼 + 𝛿)(22𝛼 2 (1 + 𝛿) − 16𝛿(3 + 2𝛿) − 𝛼𝛿(11 + 43𝛿)) 𝛽 4 −
( )
6𝛼 2 6𝛼 3 (20 + 21𝛼) + 𝛼(−216 + 𝛼(−239 + 2𝛼(50 + 63𝛼)))𝛿 − (72 + 𝛼(257 + 4𝛼(92 + 45𝛼)))𝛿 2 − (1 + 𝛼)(56 + 81𝛼)𝛿 3 𝛽 3 +
( )
𝛼 531𝛼 4 (1 + 𝛿) − 16𝛿 2 (9 + 8𝛿) + 𝛼 3 (469 + (317 − 744𝛿)𝛿) − 4𝛼 2 𝛿(223 + 𝛿(355 + 48𝛿)) − 𝛼𝛿(720 + 𝛿(799 + 319𝛿)) 𝛽 2 +
( )
2 −81𝛼 4 (1 + 𝛿) + 8𝛿 2 (1 + 𝛿) + 4𝛼𝛿(22 + 𝛿(23 + 5𝛿)) + 𝛼 2 𝛿(121 + 𝛿(197 + 12𝛿)) + 𝛼 3 (−65 + 𝛿(−34 + 111𝛿)) 𝛽 +

18𝛼 3 (1 + 𝛿) − 16𝛿(1 + 𝛿) − 8𝛼𝛿(3 + 5𝛿) + 𝛼 2 (13 + (5 − 24𝛿)𝛿).

1
Solving P(𝛽) = 0 yields two complex roots, and the two non-negative real roots, 𝛽real = 2𝛼+𝛿 and 𝛽̂. The first real root is excluded by the condition
(5),29 and the second is meaningful if 𝛿 is small enough. It follows in this case that
̂
∗ > 0for𝛽 < 𝛽.
Va∗ − Vpa

√ √
̃
28
𝛽̃′ = 𝛼(9𝛼(1+2𝛼)−17)+3(1+(𝛼−1) 1+2𝛼(−2+𝛼+2𝛼 2 ))
2𝛼(4+9𝛼(−2+𝛼+2𝛼 2 ))
and 𝛽̃′ = 𝛼(−17+9𝛼(1+2𝛼))−3(−1+ (−1+𝛼)2 (1+2𝛼(−2+𝛼+2𝛼 2 )))
2𝛼(4+9𝛼(−2+𝛼+2𝛼 2 ))
> 0.
29
In fact, according to this condition, and given 𝛿 > 0, 𝛽 > 1
2𝛼
> 1
2𝛼+𝛿
.

9
P. Pieretti, G. Pulina Economic Modelling xxx (xxxx) xxx

Appendix C. the model with immobile firms

The model developed in Section 5 can be extended in the following way. We further assume that there are m (normalized to 1) identical
multinational companies. In addition, there are n (identical) local firms that are considered as internationally immobile. Their productivity is
positively impacted by infrastructure investments. This increases home production and thus the home tax base. On the contrary, the corporate tax
rate has a negative effect.

• If the tax rate t increases, mobile firms react by shifting profit and activity in a low-tax country and thus decrease the tax base in the home
country. Similarly, the immobile firms reduce their activities, which decreases their profit.
• If infrastructure provision Gh decreases in the home country, mobile firms shift real business abroad. Similarly, immobile firms reduce their
activities, which lowers their profit.

The home tax base of a representative MNC is 𝜋im (t , Gh ) = (1 − θ(t , Gh ) − θp (t , Gh ))(1 + Gh ). The tax base of a representative immobile firm is
𝜋im (t , Gh ). As explained above, 𝜋im (t , Gh ) decreases with corporate taxation ( 𝜕𝜋𝜕tim < 0) and increases with the home country provision of public
infrastructure ( 𝜕𝜋
𝜕G
im
> 0).
h
For simplicity, we assume that the profit of an immobile firm is given by the following form that accounts for the above assumptions,
𝜆Gh Gh
𝜋im (t , Gh ) =
𝜆t t
The coefficients 𝜆Gh > 0 and 𝜆t > 0 account for the profit sensitivity of immobile firms to infrastructure provision and corporate taxes, respectively.
The welfare function of the home country can now be written as follows,

[ ] G2
Wh = t m · 𝜋m (t , Gh ) + n · 𝜋im (t , Gh ) − h .
2
Note that, setting n = 0 brings us back to the particular case of our model without immobile firms, analyzed in Section 5.
𝜆G
To guarantee an interior solution of the game, we require that n 𝜆 h < 23(𝛼−𝛼+2
1)
. After standard calculations, we get the same result as in Section 5
t
of our paper, namely that the onshore tax revenue may decrease after the complete elimination of international profit shifting if the offshore country
is efficient enough in providing attractive infrastructure (𝛽 < 𝛽̃′ ). It follows that the main conclusion of our paper is preserved when accounting for
the size of the onshore country by introducing immobile firms.
However, if the tax haven is efficient enough (𝛽 < 𝛽̃′ ) in providing attractive infrastructure, it can be demonstrated that the higher the number
(n increases) of immobile firms (or the higher the size of country H), the higher the loss of tax revenue will be after the elimination of profit shifting.
By contrast, if the tax haven is not efficient enough (𝛽 > 𝛽̃′ ), the elimination of profit shifting results in higher gains in tax revenue when the number
(n) of immobile firms is relatively higher.

References
Alms, M., Kieran, T., Cliona, D., 2016. Has BEPS signaled the death knell for U.S.
Pharmaceutical IP migration to Ireland? Tax Plann. Int. Rev. 43 May.
Bénassy-Quéré, Agnès, Gobalraja, Nicolas, Trannoy, Alain, 2007. Tax and public input
competition. Econ. Pol. 22, 385–430 April.
Bucovetsky, S., 2014. Honor among tax havens. J. Publ. Econ. 110, 74–81.
Bucovetsky, S., Haufler, A., 2008. Tax competition when firms choose their
organizational form: should tax loopholes for multinationals be closed? J. Int. Econ.
74, 188–201.
Dembour, Carole, Wauthy, Xavier, 2009. Investment in public infrastructure with
spillovers and tax competition between contiguous regions. Reg. Sci. Urban Econ. 39
(6), 679–687.
Desai, M.A., Foley, C.F., Hines, J.R., 2006a. The demand for tax haven operations. J.
Publ. Econ. 90, 513–531.
Desai, Mihir A., Foley, C. Fritz, Hines Jr., James, 2006b. A multinational perspective on
capital structure choice and internal capital markets. J. Finance 59, 2451–2487.
Desai, Mihir A., Foley, C. Fritz, Hines, James R., 2009. Domestic effects of the foreign
activities of US multinationals. Am. Econ. J. Econ. Pol. 1 (1), 181–203.
Dharmapala, D., 2008. What problems and opportunities are created by tax havens? Oxf.
Rev. Econ. Pol. 24, 661–679.
Dharmapala, D., Hines, J.R., 2009. Which countries become tax havens? J. Publ. Econ.
93 (9–10), 1058–1068.
Dischinger, Matthias, Riedel, Nadine, 2011. Corporate taxes and the location of
intangible assets within multinational firms. J. Publ. Econ. 95 (7), 691–707.
Elsayyad, May, Konrad, Kai A., 2012. Fighting multiple tax havens. J. Int. Econ. 86 (2),
295–305.
Esfahani, Hadi Salehi, Ramirez, Maria Teresa, 2003. Institutions, infrastructure, and
economic growth. J. Dev. Econ. 70 (2), 443–477.
European Commission, 2016a. COM(2016) 23 Final, Proposal for a Council Directive
Laying Down Rules against Tax Avoidance Practices that Directly Affect the
Functioning of the Internal Market.
European Commission, 2016b. COM(2016) 24 Final, Communication from the
Commission to the. European Parliament and the Council.
European Commission, 2016c. COM(2016) 25 Final. Proposal for a Council Directive
Amending Directive 2011/16/EU as Regards Mandatory Automatic Exchange of
Information in the Field of Taxation.
European Commission, 2016d. COM(2016) 271 Final, Commission Recommendation of
28.1.2016 on the Implementation of Measures against Tax Treaty Abuse.
Gonzalez, M., Schipke, A., 2011. Bankers on the beach. Finance Dev. 48 (2), 42–45.
Haufler, A., Schjelderup, G., 2000. Corporate tax systems and cross country profit
shifting. Oxf. Econ. Pap. 52, 306–325.
Hauptmeier, Sebastian, Mittermaier, Ferdinand, Rincke, Johannes, 2012. Fiscal
competition over taxes and public inputs. Reg. Sci. Urban Econ. 42 (3), 407–419.
Heckemeyer, J.H., Overesch, M., 2013. Multinationals′ Profit Response to Tax
Differentials: Effect Size and Shifting Channels. ZEW Discussion Paper, 13-045.
Hines, J.R., 2010. Treasure Islands. J. Econ. Perspect. 24, 103–126.
Hines, J.R., 2014. Policy forum: how serious is the problem of base erosion and profit
shifting? Can. Tax J. 62, 443–454.
Hong, Q., Smart, M., 2010. In praise of tax havens:International tax planning and
foreign direct investment. Eur. Econ. Rev. 54, 82–95.
Huizinga, H., Laeven, L., 2008. International profit shifting within European
multinationals. J. Publ. Econ. 92, 1164–1182.
IMF, 2015. Making Public Investment More Efficient. IMF Staff Report, Washington.
Janeba, E., Smart, M., 2006. Is targeted tax competition less harmful than its remedies?
Int. Tax Publ. Finance 10, 259–280.
Johannesen, N., 2010. Imperfect tax competition for profits, asymmetric equilibrium and
beneficial tax havens. J. Int. Econ. 84, 253–264.
Justman, Moshe, Thisse, Jacques-Francois, van Ypersele, Tanguy, 2002. Taking the bite
out of fiscal competition. J. Urban Econ. 52 (2), 294–315.
Justman, Moshe, Thisse, Jacques-Francois, van Ypersele, Tanguy, 2005. Fiscal
competition and regional differentiation. Reg. Sci. Urban Econ. 35 (6), 848–861.
Kanbur, Ravi, Keen, Michael, 1993. Jeux sans frontieres: tax competition and tax
coordination when countries differ in size. Am. Econ. Rev. 83 (4), 877–892.
Keen, Michael, 2001. Preferential regimes can make tax competition less harmful. Natl.
Tax J. 54 (4), 757–762.
Keen, Michael, Konrad, Kai A., 2013. Chapter 5 - the theory of international tax
competition and coordination. In: Auerbach, Alan J., Chetty, Raj, Feldstein, Martin,
Saez, Emmanuel (Eds.), Handbook of Public Economics, Vol. 5. Handbook of Public
Economics, vol. 5. Elsevier, pp. 257–328.
Kenny, Charles, 2009. Measuring corruption in infrastructure: evidence from transition
and developing countries. J. Dev. Stud. 45 (3), 314–332.
Morriss, A.P., 2008. The Role of Offshore Financial Centers in Regulatory Competition.
Illinois Law & Economics Research Paper No. LE07-032.
OECD, 2013a. Action Plan on Base Erosion and Profit Shifting. Organisation for
Economic Cooperation and Development, Paris.
OECD, 2013b. Addressing Base Erosion and Profit Shifting. Organisation for Economic
Cooperation and Development, Paris.
OECD, 2014. Explanatory Statement, OECD/G20 Base Erosion and Profit Shifting
Project. Organisation for Economic Cooperation and Development, Paris.

10
P. Pieretti, G. Pulina Economic Modelling xxx (xxxx) xxx

OECD, 2015. Aligning Transfer Pricing Outcomes with Value Creation, Actions 8-10 - Slemrod, J., Wilson, J.D., 2009. Tax competition with parasitic tax havens. J. Publ.
2015 Final Reports. Organisation for Economic Cooperation and Development, Paris. Econ. 93 (11–12), 1261–1270.
Overesch, Michael, 2009. The effects of multinationals′ profit shifting activities on real Stöwhase, S., 2005. Asymmetric capital tax competition with profit shifting. J. Econ. 85
investments. Natl. Tax J. 62 (1), 5–23. (2), 175–196.
Peralta, S., Wauthy, X., van Ypersele, T., 2006. Should countries control international The Economist, 2015. All My Bags Are Packed - Corporate America′s Tax-Driven Exodus
profit shifting? J. Int. Econ. 68, 24–37. Continues. The Economist, August 15th.
Pieretti, P., Zanaj, S., 2011. On tax competition, public goods provision and UNCTAD, 2013. World Investment Report 2013.
jurisdictions′ size. J. Int. Econ. 84, 124–130. UNCTAD, 2015. World Investment Report 2015.
Pritchett, Lant, 2000. The tyranny of concepts: CUDIE (cumulated, depreciated, Zissimos, Ben, Wooders, Myrna, 2008. Public good differentiation and the intensity of
investment effort) is not capital. J. Econ. Growth 5 (4), 361–384. tax competition. J. Publ. Econ. 92 (5–6), 1105–1121.

11