Content uploaded by Thomas Rixen
Author content
All content in this area was uploaded by Thomas Rixen on Jan 25, 2018
Content may be subject to copyright.
4 Double tax avoidance and tax
competition for mobile capital
Markus Leibrecht and Thomas Rixen
4.1 Introduction
Capital tax competition has received increasing attention in recent years, not
least because capital mobility, both de jure and de facto, has risen. It is fre-
quently argued that the attraction of mobile capital, other things being equal,
creates welfare gains for a country as additional tax revenues accrue and/or the
domestic capital and technology stocks increase. Thus, it is not surprising that
countries strive to increase the capital stock within their borders.
There are two largely distinct bodies of literature and corresponding research
areas in international taxation. On the one hand, there is a large and continuously
growing body of theoretical and empirical literature on tax competition (see, for
example, Wilson and Wildasin 2004; Hochgatterer and Leibrecht 2009), mainly
in the field of economics; on the other, there is a large body of mostly legal Iiter-
ature dealing with international double tax avoidance (see, for example, Lang
2003). The fact is, however, that there are only very few contributions that
address both issues at the same time (see, for example, Roin 2008). This chapter
tries to narrow the gap. Specifically, it shows why measures of double tax avoid-
ance, even though they comprise measures of bilateral tax coordination, may
actually act as instruments of tax competition.
l
The rules of double tax avoidance (DT AV) are contained in bilateral double
tax treaties (DTTs), in domestic tax rules and, in the European Union, in supra-
national tax laws. All these different levels of rules are influenced by conven-
tions promoted by international organizations. This set of rules is designed to
ensure that international investments are not over-taxed in comparison with
national investments. The goal is to achieve "single taxation" by coordinating
different countries' taxing rights so as to avoid jurisdictional overlap. However,
one consequence of the way DTAV rules are constructed and the way the
various elements interact may be "double non-taxation": sophisticated taxpay-
ers can use the rules in a way that helps them to minimize their tax payments.
They engage in so-called tax arbitrage across different jurisdictions and DT AV
rules.
However, the fact that taxpayers exploit differences in the rules in order
to minimize their tax payments is by itself not sufficient to diagnose tax
62 M. Leibrecht and T. Rixen
competition. An additional condition is that governments use taxpayers' arbi-
trage activities to attract capital investments.
2
This becomes clear if one consid-
ers that horizontal tax competition is defined as non-cooperative tax setting by
independent governments, where each government's respective policy choices
influence the allocation of mobile tax bases among the regions or jurisdictions
represented by these governments (Wilson and Wildasin 2004). This means that
independent jurisdictions interact strategically in setting taxes. Moreover, it
implies that at least three conditions must be met in order for tax competition to
exist. First, governments use tax rules intentionally to attract capital or to keep it
in the country. This condition can be met in various ways such as a discretionary
reduction in tax rates, fashioning preferential tax regimes for foreign investors or
creating (or not closing) tax loopholes. Second, capital tax bases must be sensi-
tive with respect to tax law differences, so that there is a real effect of govern-
mental actions on the allocation of mobile tax bases. Third, capital mobility must
be de jure possible and it must de facto occur (see also Bellak and Leibrecht
2008).
As has already been noted, an extensive body of empirical literature exploring
the presence of tax competition has emerged,3 but this literature focuses upon
just one tax policy instrument to attract mobile capital, namely, statutory or
effective tax rates on corporate income. Yet tax competition might occur via
more subtle channels (see also Altshuler and Grubert 2005). In particular, we
argue that there is a direct link between DT AV and tax competition. The most
obvious link is that if double taxation is curbed, the net rate of return on, say, a
foreign direct investment (FDI) is increased. This phenomenon is expected to be
particularly pronounced in the case of tax-sparing rules in some treaties between
developed and developing countries, as these rules ensure the advantages granted
by specific tax provisions. Hence, with DTAV rules in force there is an increased
propensity among firms to make foreign investments, and thus cross-border
capital flows are induced.
A more interesting link is based on the fact that DT AV rules generally rely on
legal form rather than economic substance. Thus, these rules may leave open or
create opportunities ("loopholes") for tax arbitrage, enabling businesses to shift
paper profits. Although it is likely that some of these loopholes are unintended, it
is also conceivable that governments actively use DTAV rules to tailor specific
offers to taxpayers' needs in order to attract tax base. For example, governments
can offer corporations the possibility to set up brass plate ("letterbox") com-
panies or other artificial legal entities; this encourages these enterprises to
engage in tax residence "shopping". Alternatively, governments may be reluct-
ant to include thin capitalization measures in their DTAV rules in order to make
their country more attractive for real capital investments (see, for example,
Haufler and Runkel 2008). Thus, in addition to statutory and effective tax rates,
governments may also use DTA V rules as instruments for competing over
mobile tax bases.
This chapter is structured as follows. In section 4.2, we outline the principles
and rules of international double tax avoidance and explain how they can be
Double tax avoidance 63
used for purposes of tax arbitrage. Moreover, we outline why DTAV rules can
provide an important institutional foundation of tax competition. After that, we
review in section 4.3 the empirical literature suggestive of the role of DT AV
rules in tax competition. Finally, we summarize our analysis and draw some pre-
liminary policy conclusions.
4.2 Double tax avoidance as the institutional foundation of
tax competition
In this section, we show how DTAV rules provide an institutional foundation of
tax competition. We first describe the principles and rules of DTA V that operate
on the domestic, bilateral and multilateral levels (subsection 4.2.1). We then
move on to show how the complex set of DT AV rules can be used for purposes
of tax arbitrage by taxpayers (subsection 4.2.2). Subsection 4.2.3 will then illus-
trate how governments react to taxpayers' arbitrage activities and position them-
selves in the resulting tax competition. This description also provides a stylized
historic account of the development of the international tax system.
4.2.1 Principles and rules of double tax avoidance
Double taxation is defined as the imposition of comparable taxes in at least two
countries on the same taxpayer with respect to the same subject matter and for
identical periods (Baker 200 I). It results from an overlap of jurisdiction to tax
between a residence state, where the recipient of income lives, and a source
state, where the income was generated. If both states exert their power to tax to
the full extent, the total burden on trans-border economic activities is prohibi-
tively high. In order to obtain the benefits of international liberalization, govern-
ments have a common interest in avoiding double taxation. The basic conflict is:
which country has the right to tax the income, and which country must restrict
its tax claims (see, for example, Li 2003)?
While there were some domestic provisions and a few DTTs in continental
Europe prior to World War I, after the war the problem became more pressing
when many countries put taxation on a broader basis, including general income
taxes. In the 1920s, the International Chamber of Commerce put the issue of
double taxation on the international agenda. In response to this, the League
of Nations commissioned reports by experts and convened several conferences
of technical experts and government officials. During the League years, the basic
principles and rules were developed that have guided the avoidance of double
taxation up to today. In the 1950s and 1960s, the GECD took over the position
of the League of Nations (and briefly the United Nations) as the main multilat-
eral policy forum for discussions of international tax issues (Rixen 2008: 86-99;
Picciotto 1992: 14-35).
Initially, the objective of these activities was to draft a multilateral tax treaty.
But governments persistently rejected the idea of a binding multilateral treaty.
They were nonetheless supportive of developing a model convention (Me) that
64 M Leibrecht and T. Rixen
could be employed as a template for bilateral DTTs. They insisted on keeping
the MC non-binding to allow the necessary flexibility to make nationally differ-
ing tax systems compatible with one another (Rixen 2008, 2010).
The development of the MC was marked by a conflict over the allocation of
taxing rights. Should the MC be based on the residence or the source principle?
Both principles can be normatively justified. Those emphasizing the principle of
ability to pay will favor residence taxation because the residence country is in a
better position to assess the taxpayer's worldwide income. Conversely, if one
regards the benefit principle as the more appropriate standard, this would suggest
source taxation because the source country provides infrastructure that allows
the generation of income in the first place. Both of these arguments are simple
and intuitive. None of the scholars who have discussed the issue of a desirable
allocation of taxing rights has come out in favor of one or the other principle;
rather, they have favored some solution that accords different weights to these
considerations (see, for example, Musgrave 2006; Li 2003). This normative
indeterminacy was further aggravated by a distributive corifiict between net
capital importers and net capital exporters. Exporting countries favor the resi-
dence principle, whereas importers favor source taxation. In each case, the
respective principle would result in the allocation of a bigger share of the inter-
national tax base to one or the other country (see, for example, Dagan 2000;
Davies 2004).4
Accordingly, it has not been possible to achieve a general consensus on
either principle. Instead, the solutions embodied in the various MCs that have
been developed since the 1920s (which have remained fundamentally
unchanged from that time right up to today) represent a compromise. Broadly
speaking, the primary (or exclusive) right to tax active business income is
granted to the source country (Article 7 of the OECD MC) if there is at least a
"permanent establishment" (Article 5 of the OECD MC) - that is, a fixed place
of business - in that country. The residence country, by contrast, has the
primary right to tax passive income - that is, income from financial investment
such as interest, dividends or royalties - with the source country having the
right to levy withholding taxes (Articles 10-12 of the OECD MC).s Overall, the
MC differentiates between various kinds of income and assigns each to either
the source country or the residence country. In general, the OECD MC places
greater emphasis on residence taxation.
6
For passive investment income, the
rates for royalties are different from those for dividends or income. The OECD
MC is thus based on a schedular system of taxation. The residence country is
obliged to provide relief from double taxation in cases of full or limited source
taxation. This can be done by granting either a foreign tax credit for the tax
paid at source on the tax due in the home country or a full exemption of that
income from home tax. In addition, bilateral treaties contain provisions on
information exchange between tax administrations (Article 26 of the MC). This
is the only provision of DTTs that is not primarily concerned with the avoid-
ance of double taxation but focuses instead on the problem of potential tax
evasion and avoidance.
Double tax avoidance 65
In parallel to the drafting of the non-binding multilateral MC (and in some
cases even prior to it), governments have developed domestic tax rules to avoid
double taxation (Rixen 2008). Basically, all countries provide for relief of double
taxation unilaterally. Their respective national tax codes contain provisions for
exempting foreign income from taxation, crediting the taxpayer or granting a
deduction (partial relief) for taxes paid on such income in the source country. In
addition, governments conclude bilateral OTIs on the basis of the MC. Basi-
cally, all of the more than 3,000 bilateral tax treaties adopt the basic structure
and the provisions of the MC (Owens 2008). Thus, most OTTs are similar in
terms of the topics covered, their structure and even the language used. Seventy-
five percent of the wording in any given OTT will be identical to that of any
other OTT (A vi-Yonah 2007). Often, technical innovations that come up in
bilateral treaties or in domestic provisions are integrated into the multilateral MC
so that experience gained at the national and bilateral levels can be disseminated
multilaterally to all parties. Other innovations are initiated by the DECO, which
constantly strives to modernize and adapt the Me. Through this interaction of
domestic, bilateral and multilateral activities, a common understanding of double
tax avoidance is developed.
7
In other words, the complex of rules which we sum-
marily refer to as OTAV rules consists of domestic rules of international taxa-
tion and of binding bilateral OTTs, both of which are informed by the principles
developed within multilateral policy forums, most importantly the DECO (Rixen
2010).
In line with the interests of governments, the particular solution embodied in
the MC is sovereignty-preserving - that is, it ensures that countries are as free as
possible to apply their own national tax laws (Rixen forthcoming). The MC
defines a series of legal constructs intended to allocate the right to tax among the
jurisdictions involved (see, for example, Bird 2000). For example, the concept of
a permanent establishment (PE) codifies what is taxable as a separate entity of a
multinational enterprise (MNE) in the country of source. This and other con-
structs are chosen in such a way as to interfere as little as possible with national
tax laws. They merely allocate the right to tax to governments, without prescrib-
ing whether or how they ought to exercise this right. Governments retain full
authority to design all elements of their respective tax regimes - namely the tax
base, rate and system - independently of other governments. The idea is that of
territorial disentanglement of different tax systems. There is no harmonization
of national tax systems, but only coordination of different regimes. In this sense,
the term "international tax" is a misnomer, since there is no overriding interna-
tional law of taxation, but only rules of allocation that operate at the interfaces
of different national tax systems (Li 2003). Emblematic of the sovereignty-
preserving setup of international taxation are the rules for allocating the profits
of MNEs to the various countries in which the MNE operates. Under the "sepa-
rate entity approach", the allocation should be the same as would result if the
different entities of a multinational group were independent actors transacting on
a market. This so-called arm's-length standard (ALS) makes it unnecessary to
agree on a common tax base.
66 M Leibrecht and T. Rixen
Thus, while the interfaces of different tax systems in the form of DTAV rules
have been harmonized to a considerable extent - even though important differ-
ences remain among bilateral tax treaties - national tax rules remain dissimilar.
4.2.2 How double tax avoidance enables tax arbitrage
Although rules of double tax avoidance explicitly address only governments and
tell them how to structure their international tax relations, the same rules (implic-
itly) pre-structure taxpayers' avoidance or planning strategies. There are several
techniques of tax arbitrage, which is defined as "tak[ing] advantage of inconsist-
encies between different countries' tax rules to achieve a more favourable result
than that which would have resulted from investing in a single jurisdiction"
(Shaviro 2002: 319).
First, the manipulation of arm's-length transfer prices is a method of avoid-
ance that builds on an important aspect of the DTT regime. According to Article
7 (2) of the OECD MC, the profits of an MNE have to be attributed to the differ-
ent affiliates as if their transactions with each other were those of independent
enterprises exchanging their products and services at regular market prices.
However, it is often difficult to determine the "correct" market prices, since there
is no regular market on which the goods in question are traded.8Accordingly,
taxpayers enjoy legal maneuvering room to set transfer prices in a way that leads
to the allocation of larger shares of profit to low-tax countries.
A second tax planning technique is treaty shopping - that is, the "accessing of
treaty benefits by persons who are not resident of either treaty state through the
use of an entity that qualifies as a resident of one of the states" (Li 2003: 106).
For example, if a company wishes to get access to cheap capital from residents
of a country with which there is no tax treaty, it may be advisable to let a finan-
cial subsidiary in a treaty partner country issue the bonds (see, for example,
Papke 2000).
Third, the schedular structure of DTTs allows taxpayers to reclassify their
financial flows in a tax-optimal way. Most importantly, they can substitute
debt for equity. Very often, this reclassification is facilitated by the possibility
of hybrid financing and hybrid entities, which we discuss here in more detail
for three reasons. First, the use of hybrid financial instruments provides a vivid
illustration of how the structure of DTTs can be utilized to reclassify financial
flows tax optimally. Second, hybrid financing can be interpreted as a sophistic-
ated way to substitute debt for equity with the intention of circumventing thin
capitalization rules (see the following subsection, subsection 4.2.3) - in many
cases without actually increasing the debt to equity ratio. Third, although
hybrid instruments may be issued for a variety of non-tax reasons (Duncan
2000), taxation issues do have considerable impact on management's financing
decisions.
What makes hybrid instruments such a potent tool in tax planning is the
simple fact that in the majority of countries, financial instruments must be treated
either wholly as equity or wholly as debt for fiscal purposes.9In other words,
Double tax avoidance 67
income is treated either as a distribution of profit (or dividends, in a broad sense),
or as interest. This in tum is crucial for two reasons: first, the classification gen-
erally determines whether or not the issuer can treat the payments as tax deducti-
ble; second, in many cases it determines whether the income from the
instruments is tax-exempt in the hands of the recipients (Eberhartinger 2005;
Helminen 2004; Six 2007; Wittendorff and Banner-Voigt 2000). Only in rare
cases is the fiscal treatment of an instrument split.
De facto, a wide variety of financial instruments incorporate elements of both
equity and debt.
1O
Because these financial tools usually cannot be classified
clearly as either one or the other, they are referred to as "hybrid instruments", or
collectively as "mezzanine financing". The spectrum of hybrid instruments
ranges from corporate shares with features typical of loans (such as certain pref-
erence shares) to loans with features usually associated with equity investments
(such as participation in profits and losses). Such equity-type loans include,
among other things, jouissance rights, silent partnerships, participation bonds,
convertible bonds, warrant bonds, profit-participating loans and preference
shares. For fiscal authorities, the wide variety of hybrid instruments in use and
their rapid evolution makes an unambiguous classification of any given hybrid
instrument all but impossible. The need to apply general criteria in the classifica-
tion of hybrid instruments opens up significant opportunities for tax planning.
Evidently this is even more important for cross-border transactions. Hybrid
instruments are used effectively as flexible, tailor-made forms of finance because
in cross-border situations both countries involved must decide on the classifica-
tion of the instruments and could, in some cases, decide differently. Thus, if a
payment is deductible as interest in the source state and exempt as remuneration
of equity (dividend) in the parent company's state of residence, the result will be
double non-taxation. In the opposite case, where the hybrid instrument is treated
as equity in the source state and as debt in the state where the parent company
resides, the difference in classification may result in double taxation, with the
payment subject to withholding tax in the source state and to income tax in the
parent company's state of residence. This is where the DTTs should step in to
prevent double taxation.
The relevant distributive rules in DTTs for the treatment of income derived
from hybrid financial instruments are Articles 10 (on dividends) and 11 (on
interest). Articles 10 and 11 of the OECD MC give the residence state an unlim-
ited right to tax the income received; at the same time, these rules do not deny
the source state a right to tax that same income by levying a withholding tax. To
avoid double taxation, the OECD MC therefore obliges the state of residence to
credit the taxpayer with the tax withheld by the source state against the corporate
tax that the former desires to levy. Articles 23A and 23B of the OECD MC
provide for the use of the standard credit method with a per-country limitation in
these cases.
The OECD MC curtails the rights of the source state by limiting the amount
of withholding tax it may levy: Article 10 limits it to 5 percent in the case of
associated companies
l!
and to 15 percent in all other cases; Article 11 limits the
68 M Leibrecht and T. Rixen
permitted withholding tax to 10 percent. These limits are the subject of bargain-
ing in treaty negotiations, and individual conventions may differ considerably
from the OECo MC.12 Similarly, it is also possible for a specific DTT to deny
the source state the right to levy withholding tax; this is more frequently the case
with interest payments than with dividends.13 Whereas the actual levels of with-
holding tax provided for in a specific OTT are relevant in practice, it is important
to point out that oTTs may give the source state the right to levy withholding
tax on both dividends and interest, on neither of them or - most importantly - on
only one of them. This last case is clearly important for hybrid finance because
the classification of the return as dividend or interest then determines whether
withholding tax can be levied or not.14
In the standard case where both countries apply the same classification criteria
to decide whether the return on a certain hybrid instrument falls under Article
10
or Article 11, it is merely a matter of tax planning to choose hybrid instruments
whose returns fall under the more favorable of the two articles. More difficult to
resolve is the situation where a OTT provides for a withholding tax in accordance
with only one of the articles but the countries concerned do not apply the same
article to the return of a particular hybrid instrument. In one particularly problem-
atic case, the source state (country X) applies the article that provides for with-
holding tax (say, the dividend article) and levies the appropriate tax. The
recipient's state of residence (country V), on the other hand, applies the article
that does not provide for withholding tax. Must Y then credit the taxpayer with
the tax withheld in X, even if according to Y's interpretation of the treaty no such
withholding tax is permitted?15 Although the prevailing German legal opinion
(see, for example, Wassermeyer 2004) in part seems to suggest that the state of
residence is not obliged to credit the taxpayer in this case, which would result in
double taxation of the return despite the OTT, the issue remains unresolved. It
seems reasonable to assume that some countries will credit the withholding tax
while other countries will decline to do so in these circumstances. In any event,
this problem makes it clear why the application of the two relevant articles is
important when hybrid instruments are used, particularly if the accepted view is
that the OTT must be interpreted autonomously (Lang 1993, 1998, 1999).
The essential criterion for distinguishing between these two distributive rules
is the existence of a corporate right. The yield of a hybrid instrument qualified as
a corporate right under Article lOis defined as a dividend in accordance with the
purposes outlined in that article. In all other cases, the yield constitutes interest
under Article 11. In order for an investment to qualify as a corporate right under
the OECo MC, the investor must accept the risk of possible loss of the invest-
ment in a way comparable to the risk assumed by a regular shareholder. Accord-
ing to the prevailing German doctrine, this is the case if the investment involves
participation in both the profits and the liquidation proceeds of the issuing
company. Only if these two conditions are satisfied is the investor's risk of loss
comparable to that of a regular shareholder.16
The OECo MC, however, does not explicitly mention these conditions or
indicate how they can be satisfied for purposes of qualifying a financial
Double tax avoidance 69
instrument as equity. Hybrid financial instruments have numerous possible
characteristics regarding participation in entrepreneurial risk, which makes it dif-
ficult to determine whether these conditions are satisfied in a particular case.
This in tum will inevitably lead to situations where two contracting states disa-
gree on whether the characteristics of a hybrid instrument are such as to qualifY
that instrument as a corporate right and thus as generating dividends. It is there-
fore possible that, in the situation of a particular taxpayer and a given form of
hybrid finance, one state applies the dividend article (Article 10) while the other
applies the interest article (Article 11). Depending on the relevant tax treaty, this
might result in double taxation or double non-taxation of the taxpayer's
incomeY From the perspective of tax planning this situation holds considerable
potential because it implies that a group in a given set of countries may be able
to choose an appropriate hybrid instrument that allows for double non-taxation,
thus creating in fact tax-free income. An enterprise or business group with par-
ticular financial needs would be able to choose appropriate countries for its sub-
sidiaries in order to optimize or even eliminate the tax burden on payments
received (treaty shopping).
With these and other similar techniques, taxpayers make sure that profits are
taxable in low-tax countries, while losses occur in high-tax states. What these
methods have in common is that none of them relies on the relocation of real
business activities, but rather on the mere shifting of paper profits and losses.
This form of tax arbitrage is possible because the OTT regime gives states the
freedom to design their own national tax laws. The sovereignty-preserving
approach to double tax cooperation creates an opportunity structure for taxpay-
ers to engage in tax arbitrage.
4.2.3 D TA V rules as instruments of tax competition
The sovereignty-preserving character of DTAV rules creates not only an oppor-
tunity structure for taxpayers but also one for governments to engage in tax com-
petition. This manifests itself in different ways. To begin with, since DTTs may
reduce the overall tax burden on international investment but at the same time
contain provisions that increase investors' legal certainty, governments may
want to conclude such treaties in order to signal the attractiveness of their
respective countries as a location for international investment (see, for example,
Dagan 2000). Whereas this would make DTTs important instruments in the com-
petition for mobile capital, clauses on the exchange of information might reduce
the possibilities for tax avoidance, which could hamper FDI.
But there are other ways to use DTAV rules in tax competition. Most import-
antly, tax haven governments make special offers to satisfY taxpayer demand for
tax minimization. This is possible because DTA V rules set no limits on how
governments can design their national tax legislation and because, in general,
legal form and not economic substance is sufficient for establishing tax resi-
dence. Also, governments actively design solutions that can help taxpayers to
play out the differences among different DTTs or different domestic laws and
70 M Leibrecht and T. Rixen
DTTs. This is most easily visible if one considers the role tax havens play in the
tax competition game. Tax-haven governments offer very low or zero tax rates
in combination with tailored solutions, making it easy for businesses to establish
legal tax residence under their jurisdiction. One way of doing this is to allow the
incorporation of brass plate businesses which serve no substantive economic
purpose, merely the tax-privileged holding of assets for a multinational group.
Tax havens actively search for market niches and try to specialize in different
tax planning activities.
ls
This illustrates that tax havens are in competition with
each other and with high-tax countries for the assignment of so-called paper
profits. This also implies that high-tax countries come under competitive pres-
sure; they must take the threat of the erosion of their tax base into consideration
in the design of their tax systems.19
Many high-tax countries reacted to tax arbitrage with a combination of trying
to curb it and efforts to remain competitive vis-a-vis other countries. With
respect to curbing tax arbitrage, governments have developed unilateral meas-
ures against tax avoidance. One example is the so-called controlled foreign com-
panies (CFC) legislation, which targets the use of foreign subsidiaries as base or
conduit companies in tax havens. In the 1960s, the United States was the first
country to introduce unilateral anti-avoidance legislation. Resident shareholders,
for example the parents of an MNE, controlling a subsidiary in a tax haven are
taxable on the passive income (e.g. interest incurred from the lending of liquid
funds) of the subsidiary in the current period, regardless of whether that income
is actually distributed to them or not. Thus, CFC rules stretch the original DTAV
principle of treating all parts of an MNE as separate national entities20 and in so
doing pierce the "corporate veil" of the tax haven entity. In the meantime, such
rules have diffused via the OECD to all major capital-exporting nations (see, for
example, Arnold 2000).
Another example of an attempt to curb tax arbitrage is so-called thin capitali-
zation rules. These rules identify a threshold of excessive debt on the basis of the
debt-equity ratio. The rules only apply to non-resident lenders holding a signi-
ficant percentage of the shares of the resident corporation.
2I
If the criteria are
met, then the debt is treated as "hidden equity" and the interest payment is not
deductible.22
A final example of an area in which efforts at curbing abuse have been made
is transfer pricing. Again the United States was the first country to push for
changes. In the I990s, it adopted new transfer pricing guidelines; OECD recom-
mendations were subsequently also reformed. Since then, the guidelines have
been in a state of almost perpetual adaptation. Overall, the new guidelines move
the actual rules closer to considering the consolidated profits of the MNE but
take great care to formally reinforce the principle of separate entity accounting
on a transactional basis. With the introduction of advanced pricing agreements
(APAs) in many countries and their promotion by the OECD, this trend has
become even more pronounced.
23
However, the effectiveness of these measures is questionable. In the case of
transfer prices, for instance, the manoeuvring room for taxpayers remains high
Double tax avoidance 71
because the range of permissible prices is large. At the same time, the procedure
imposes serious administrative costs on the taxpayer. As for the unilateral anti-
avoidance measures, what can be observed is that they generally target certain
circumscribed tax arbitrage schemes which taxpayers can simply circumvent by
designing more sophisticated schemes. For example, CFC rules are often side-
stepped by creating hybrid entities.24
What is observed in practice is a kind of proliferation spiral. Taxpayers react
to anti-avoidance rules by devising more sophisticated techniques of tax arbi-
trage; these in tum are then targeted by new anti-avoidance rules by govern-
ments, and so on. In effect, governments continue to react to taxpayers'
arbitrage activities. A likely reason why governments do not design broad and
effective rules against tax arbitrage is that they are under competitive pressure
vis-a-vis other high-tax countries which are willing to leave loopholes open for
their taxpayers.25 One recent example of this is the introduction of a so-called
Zinsschranke (i.e. a unilaterally binding boundary on debt financing of affili-
ates) in German corporation tax law. Originally, the finance minister planned to
introduce a thin capitalization rule that foresaw the non-deductibility of 50 per
cent of interest expenses. However, this proposal met with opposition from
business interests and parliamentarians of the Christian Democratic Union
(CDU), one of the parties making up the coalition government. Eventually, the
Zinsschranke passed by the government only foresees the non-deductibility of
30 per cent of net interest expenses (those expenses above interest receipts),
and this only if none of three exceptions (small company clause, group clause
and escape clause (see Homburg 2007: 605) is applicable (Jarass 2007). Very
recently, following continuous lobbying of business interests against the rule,
the small company clause has been extended as part ofthe government's stimu-
lus package against the global recession: instead of the rule being applied to
interest expenses above €l million, the threshold has been increased to €3
million.
To summarize, DTAV rules not only achieve the avoidance of double taxa-
tion but also provide an institutional foundation of tax arbitrage (for taxpayers)
and, further, DT AV rules become instruments of tax competition (for home and
host-country governments).26 Some evidence for a role of DTA V rules in tax
competition is presented in the next section.
4.3 Empirical evidence on the role of DT
A V
in tax
competition
In section 4.2, we outlined several channels through which DTA V rules can
result in tax minimization by MNEs and thus can be used for tax competition by
governments. This section tries to assess the empirical relevance of these
channels. Unfortunately, no direct evidence is available, underscoring the role of
DTAV rules in tax competition for mobile capital, so we must rely on indirect
evidence here. Following Griffith and Klemm (2005), direct evidence for tax
competition is based on empirical models relating a country's tax rate on
72 M Leibrecht and T Rixen
corporate income to the weighted tax rate of competitor countries (i.e. on
reaction-function approaches). The seminal paper in this respect is Devereux
et
al. (2004). A direct study would, thus, relate rules of DTAV by one country to
that of competitor countries.27 However, the papers surveyed can provide indi-
rect evidence for the role of DT AV rules in tax competition as they focus on a
precondition for tax competition to set in, namely the sensitivity of mobile
capital (especially FDI and paper profits) with respect to tax conditions.
Specifically, we survey (i) empirical papers about the effect of DTTs on
FDI, (ii) papers on the relevance of tax-sparing rules for FDI to developing
countries, and (iii) studies dealing with the empirical relevance of profit shift-
ing, whereby a particular focus is put on transfer pricing and the impact of thin
capitalization rules on debt financing of foreign affiliates. Finally, we present
(iv) the main results of a study dealing explicitly with the role of "hybrid
entities".
4.3.1 Preliminary remarks
The papers surveyed are structured along several dimensions (main outcomes,
databases used, estimators applied, etc.).28 A particular focus is put on the deri-
vation of the implied semi-elasticities. Semi-elasticities represent the percent-
age change in the endogenous variable (FDI, pre-tax profit, etc.) when the
exogenous variable of main interest (statutory tax rate on corporate income,
dummy variable for DTT in force or not, etc.) changes by one unit (e.g. a
change of 1 percentage point in the tax rate or the change of a dummy variable
from
0
to 1). The derivation of semi-elasticities is intended to make the results
of the heterogeneous studies surveyed somewhat more comparable (see also De
Mooij 2005).
The empirical models used in the papers do not usually lead directly to semi-
elasticities. Thus, the regression coefficients shown in the papers need to be suit-
ably transformed in a first step. The empirical models applied in the various
papers are usually based on "level-level" or on "log-level" specifications. Level-
level specifications relate the endogenous variable in levels (e.g. FDI in millions
of euros) to the independent variable of main interest, also measured in levels
(e.g. the statutory tax rate on corporate income measured as a percentage or a
dummy variable indicating the presence of a DTT). In contrast, log-level speci-
fications have the dependent variables in logarithmic form (e.g. 10g(FDI)). The
advantage of the latter specification is that the implied semi-elasticity is made
directly visible by multiplying the regression coefficient by 100.29
In the case of level-level specifications, the implied semi-elasticity can be
deduced by:
e(s) = al(y_mean) x100 (1)
with abeing the regression coefficient and y_mean the sample mean value of the
endogenous variable.
Double tax avoidance 73
Moreover, some empirical models are based on elasticities (e) as regression
coefficients ("log-log specifications"). In these cases the implied semi-elasticities
are derived as:
e(s)
=
e/(x_mean) x100 (2)
with
e
=
elasticity
=
regression coefficient and x_mean being the sample mean
value of the independent variable of main interest
Not all papers provide the information necessary (e.g. the mean values of y
and x) to derive the semi-elasticities from the regression coefficients. Therefore,
only papers providing this information are included in the survey.
4.3.2 A brief summary of empirical studies
4.3.2.1 Double taxation treaties and foreign direct investment
Empirical papers dealing with the impact of DTTs on FDl are quite scarce. This
is even more so for studies analyzing the impact of tax-sparing rules contained
in DTTs on FOI to developing countries. Basically, signing a OTT might on the
one hand lead to more FDl because double taxation of income is avoided and
uncertainty is reduced. On the other hand, however, DTTs might also reduce the
possibilities for tax evasion and avoidance (Egger et al. 2006). Thus, empirically
speaking, the effect of OTIs on FOI is ambiguous a priori.
Concerning studies dealing with FOI and DTT,3° the early paper of Blonigen
and Davies (2000) concludes that the signing of DTTs does not have relevant
effects on the volume ofFOI because most regression coefficients lack statistical
significance. Thus, a semi-elasticity of zero is plausible. In a follow-up study
(Blonigen and Davies 2002), the same authors find that FOI is lower if a new
treaty is signed. This suggests that the possibilities for evading taxes via FOI are
fewer if a viable tax treaty is in force.3' The analyses of Blonigen and Davies are
based on level-level specifications. In more recent work (Blonigen and Davies
2004b), the same authors argue in favor of using a "log-level" specification.
Applying the latter leads the authors to conclude that new treaties do not seem to
impact positively or negatively on FDI because the regression coefficients are
not statistically different from zero. To sum up, the papers of Blonigen and
Davies imply that DTTs do not impact on the level of FDI.
However, two aspects of the Blonigen and Davies studies are worth mention-
ing. First, they rely largely on data for developed economies; and second, they
consider the signing of a OTT to be an exogenous event. Concerning the first
aspect, Neumayer (2007) provides evidence that the presence of a OTT increases
FOI in middle-income developing countries; thus, he finds a positive impact of
DTTs on FDI. Egger et al. (2006) allow the signing of a OTT to be an endog-
enous event. Specifically, it is more likely that countries will sign a OTT in cases
where bilateral FOI is already present. Using a rather sophisticated two-step
approach, they find that the signing of a new OTT reduces FOI significantly.
74 M Leibrecht and T. Rixen
This result is consistent with the view that DTTs effectively reduce the possibil-
ities to avoid taxes via FDI.
If we tum to recent studies on the topic, Barthel
et al.
(2008) conclude that
DTTs lead to substantial and significantly higher FDI stocks. The distinguishing
feature of this study is the usage of a large dataset containing both developed
and developing countries. Finally, Davies et al. (2009) again find little evidence
for an effect ofDTT on the level of total sales acting as proxy for foreign invest-
ment. However, they do find evidence supporting the contention that treaty for-
mation has a positive impact on the probability of new investment in a country.
Thus, tax treaties seem to have no effect on the volume of investment (given that
a firm has already invested in a country), but they do seem to have an impact on
the location of the investment.32
So what can we conclude from these papers with respect to our assertion that
DTAV rules create an institutional foundation of tax competition? Since the
more recent studies find evidence that DTTs positively impact FDI (especially to
developing countries) and new investments, then one precondition for tax com-
petition to set in - namely, the sensitivity of FDI to changes in the taxation
environment - seems to be fulfilled. Put differently, without DTTs, less capital
would move across borders, and tax competition for FDI between countries
would be of less importance.33 However, one must bear in mind that the results
of Egger et al. (2006) are evidence against our assertion. Their results imply that
DTTs are effective in reducing the possibilities for MNEs to avoid taxes and that
DTTs are not instruments of tax competition for governments. To sum up, the
available literature supports our assertion if developing countries are included in
the sample and if the impact of DTTs on new investments is considered.
4.3.2.2 Tax-sparing clauses andforeign direct investment
Tax-sparing clauses are sometimes included in DTTs signed by developed coun-
tries, in particular if applying the credit system in international taxation (e.g. Japan
or the United Kingdom), and developing countries. These rules have the aim of
allowing foreign investors to partake of the special tax advantages provided by the
host countries ofFDI. Put differently, if a tax-sparing clause is included in a DTT,
then the home country of FDI calculates the residual tax liability of the investor as
if the special tax advantage were not provided by the host country (see Hines 1998).
Only two empirical papers on this issue are available so far. Both are based
on Japanese FDI data and both find a huge impact of tax-sparing rules on FDI
(see the semi-elasticities contained in Appendix A2, p. 86). This evidence is con-
sistent with the results of Neumayer presented above, which show a significant
impact of DTTs on FDI to developing countries. Thus, the studies surveyed
provide fairly solid evidence for the positive impact of DTTs on FDI to develop-
ing countries, especially if they include tax-sparing clauses. Hence, overall the
evidence is in favor of DTTs making capital mobile across borders. Thus, DTAV
rules are capable of spurring tax competition, which supports our assertion that
these rules are an institutional foundation of tax competition.
Double tax avoidance 75
4.3.2.3 Tax-motivated profit shifting
As has already been outlined, various means to shift profits from high- to low-
tax countries exist. These include transfer pricing, debt financing of affiliated
companies or the location of intangible assets. Profit-shifting incentives are
exerted by differences in statutory tax rates on corporate income across coun-
tries (see, for example, Weichenrieder 2009). These differences lead to tax-
motivated changes in the location of pre-tax profits via the various means to
shift profits.
The papers surveyed here are separated into three categories: (i) papers
dealing with profit shifting in general, (ii) papers dealing with profit shifting via
transfer pricing or the location of intangible assets, and (iii) papers dealing with
the effect of thin capitalization rules on debt financing offirms.
The papers summarized in Appendix A3 relate a proxy for a firm's declared
profits (pre-tax or net of tax, depending on the study) to the relevant (home or
host country) statutory tax rate on corporate income.34 De Mooij's paper (200.5)
also includes a survey of empirical studies. The author calculates a semi-
elasticity of -2, meaning that a I percentage point shift in the relevant tax rate
changes declared profits by about 2 percent. More recent studies (Weichenrieder
2009; Huizinga and Laeven 2007) broadly confirm this evidence even if the
implied semi-elasticities are somewhat lower (about + 1 and -1.30).35
Concerning our assertion, these results imply that profit shifting is an empiri-
cally relevant phenomenon and that DTAV measures are not effective in curbing
it. However, not curbing profit shifting might be in the interest of governments
because this could lead to an inflow of tax revenues (in a low-tax country) or
because profit-shifting opportunities probably make real capital less sensitive to
differences in statutory tax rates across countries.36
The papers whose results are shown in Appendices A4 and A5 shed more
light on the various channels of profit shifting (transfer pricing, location of intan-
gibles and thin capitalization of foreign affiliates). Clausing (2003) was among
the first to address the transfer pricing issue directly. Clausing relates the export
and import prices of US multinational firms to the statutory tax rate on corporate
income of the host country of a foreign affiliate. She finds substantial evidence
for tax-motivated transfer pricing in US intra-firm trade prices. Specifically, the
lower a host country's tax rate, the lower are US intra-firm export prices and the
higher are US intra-firm import prices relative to corresponding arm's-length
prices. Based on Clausing's results, semi-elasticities of about +4 for export
prices and -4.7 for import prices are derived (cf. Appendix A4).
Bernard
et al.
(2008) also give direct evidence for the importance of tax-
motivated transfer pricing. Their paper shows that the wedge between the arm's-
length export price and the intra-firm export price of US-based multinationals is
negatively related to the taxation of corporate income in the foreign import
country. This response indicates that tax-motivated transfer pricing is present in
the intra-firm export prices of US-based multinational companies. The mean
semi-elasticity derived is -1.25. This value implies that a 1 percentage point
76 M Leibrecht and T. Rixen
increase in the host country's tax rate on corporate income leads to a decrease in
the wedge between arm's-length export prices and intra-firm export prices of
1.25 percent.
Dischinger and Riedel (2008) analyze the tax sensitivity of the location of
intangible assets with respect to the statutory tax rate on corporate income.
They show that the location of holdings and investments in intangible assets is
indeed inversely related to the statutory tax rate on corporate income. This rela-
tionship can be considered as indirect evidence in favor of tax-motivated trans-
fer pricing in the intra-firm trade of intangibles. The derived semi-elasticity
is
-1.20.
37
Concerning our assertion, these results collectively imply that profit shifting
via transfer pricing is an empirically relevant phenomenon and that stipulations
concerning transfer pricing contained in DTA V rules are by no means effective
in curbing it. However, as in the case of profit shifting in general, it might be in
the interest of governments not to inhibit profit shifting through transfer
pricing.
Turning to papers dealing with the impact that thin capitalization rules exert
on the financing structure of affiliated companies, we see that such rules do
indeed have a significant effect on the level of debt financing (cf. Appendix A5).
The results indicate that the introduction of thin capitalization rules has a pro-
nounced effect on the debt-to-asset ratio of an affiliated company. This means
that governments could reduce the amount of profit shifting via debt financing
with relative ease.38 But even in the European Union, as of 2007 seven Member
States had not applied thin capitalization regimes. Worthy of note here is that
among those countries which have not applied such rules are many smaller
nations that are known to be among the more aggressive tax competitors (see
Haufter and Runkel 2008; Dourado and de la Feria 2008). Furthermore, the rules
in operation in the other EU member countries vary according to the method
adopted, the scope of application and their effects (see Dourado and de la Feria
2008). Not applying thin capitalization rules might be considered as an effort by
governments to attract foreign real capital (see also Altshuler and Grubert 2005;
Haufter and Runkel 2008), most likely coupled with a positive effect on the
domestic capital stock and with positive technology spill-over effects.39 Again,
these results favor our conjecture.
4.3.2.4 The role afhybrid entities
As is suggested above (in section 4.2), tax minimization using hybrid entities
or hybrid securities is likely. But empirical evidence on the role of hybrid
instruments is scarce. One notable exception is the paper of Altshuler and
Grubert (2005), who analyze the effects of a specific amendment to CFC rules
in the United States in 1997.
In
particular, new regulation in 1997 introduced
"hybrid entities". Altshuler and Grubert describe these as "business operations
that are regarded as corporations by one country while being an unincorpo-
rated branch to another" (ibid.: 5). Hybrid entities allow US companies to
Double tax avoidance 77
avoid the current tax under the CFC rules on intercompany payments (e.g.
interest and royalties) made from an affiliate in a high-tax country (H) to
another affiliate in a low-tax country (L). At the same time, country H also
allows such payments to be deductible. The new regulations are usually
referred to as "check-the-box" rules because they give firms the possibility
simply of checking a box on a tax form to either identify an entity as a separate
corporation or classify it as the branch of another corporation (see Altshuler
and Grubert 2005 for details). The introduction of these rules in 1997 provides
sufficient varience to distill the empirical importance of hybrid entities. On the
basis of this change in US national tax law, Altshuler and Grubert provide
evidence that MNEs could save huge amounts of taxes by using hybrid
entities.
One piece of evidence in this respect is based on the comparison of the tax-
rate sensitivity of declared pre-tax profits by US foreign affiliates with respect to
the statutory tax rate on corporate income in the host country before and after the
introduction of check-the-box. Altshuler and Grubert show that the tax-rate sen-
sitivity in 2000 (i.e. after the introduction of check-the-box) was higher in abso-
lute value than it was in 1996 (the year before the introduction). The
corresponding semi-elasticities given in Table A6 are -1.48 for 1996 and -2.07
for 2000, respectively. This difference in the magnitudes of the two mean values
is suggestive evidence in favor of an increase in tax-motivated profit shifting via
hybrid entities after the 1997 rule change. Specifically, according to Altshuler
and Grubert, it is possible that highly profitable subsidiaries in high-tax countries
became part of a consolidated hybrid entity based in a tax haven or that "high-
tax host countries may have reacted to the increasing tax sensitivity of invest-
ment by easing up on their transfer pricing and thin capitalization rules in order
to attract mobile corporations" (2005: 17).
Another piece of evidence contained in Altshuler and Grubert which is con-
sistent with the increasing importance of hybrid entities is that "there are less
CFCs in the 2000 sample than in 1996" (2005: 16 footnote 3). Moreover, the
authors find that the total tangible capital in all locations grew by 28 percent
from 1996 to 2000. But in five low-tax countries, tangible capital grew by almost
200 percent. According to Altshuler and Grubert, most of this growth "reflects
the consolidation of the low-tax CFC with the operations of a high-tax affiliate"
(ibid.: 18).
Concerning our assertion that DT AV rules provide an institutional foundation
of tax competition, the example of the 1997 check-the-box change in US CFC
rules suggests that unilaterally applied measures (in this case by the home
country of FDI) make it easier for MNEs to avoid taxes. The introduction of
check-the-box can also be considered as a home-country instrument to attract -
or at least to avoid losing - firms because the ability to hide profits makes it
more likely that a company will not opt to relocate its headquarters.
To sum up, the indirect evidence provided in this section suggests that DTAV
rules do indeed serve as an institutional foundation of tax competition: DTTs seem
to increase the propensity offirms to become multinational via FDI. DTAV without
78 M Leibrecht and T. Rixen
thin capitalization rules, despite their effectiveness in curbing profit shifting, can be
seen as an example for host countries using not only statutory tax rates to compete
for mobile capital but also more subtle channels. The presence of tax-motivated
transfer pricing via intra-firm trade or through the location of intangibles, despite
the existence of DTTs and transfer pricing regulations, is another example in this
respect. The check-the-box regulation represents an example for the role of unilat-
eral DTAV rules introduced by a home country as instrument of tax competition
(for the latter, see also Altshuler and Grubert 2005).
4.4 Summary and conclusions
This chapter explores the role DTAV rules play in tax competition for mobile
capital. We first sketched various channels, introduced by DTAV rules, through
which taxpayers can minimize tax payments and which also constitute channels
for governments to compete for mobile capital. Second, the overview of existing
empirical research provides indirect evidence for our assertion that tax competi-
tion via subtle DTAV-related channels exists. We may conclude that DTAV
rules playa prominent role in the tax competition game. The deeper reason for
this lies in the patchwork structure of DTAV rules. The fact that there are differ-
ences among the various domestic and bilateral rules creates possibilities for tax
arbitrage, which in tum creates tax competition among countries and thus rein-
forces the differences among the tax rules.
What policy conclusions follow from these findings? Specifically, since
DTAV rules also comprise measures of tax coordination (e.g. DTTs), what con-
clusions can then be derived concerning the discussion on the virtues and pitfalls
of tax coordination and tax harmonization?
The answer depends on the welfare effects of tax competition. If one sees tax
competition as a process to restrain selfish governments or to induce policy inno-
vations, then the answer would probably be "do nothing because there are no
virtues of tax coordination". However, if one sees tax competition as a process
leading to welfare decreases, then it is clear that the current sovereignty-
preserving approach to tax coordination is insufficient to address the issue of tax
competition effectively. As long as governments remain free to determine central
aspects of their respective tax systems independently, they will always have an
incentive to tailor their tax regimes to the specific needs of taxpayers wishing to
optimize their international tax payments. Neither unilateral measures (which
governments will often also use as strategic instruments in tax competition) nor
bilateral tax treaties can be effective instruments to address the inherently multi-
lateral problems of tax arbitrage and tax competition.
What would be needed instead is tax coordination by delegation40 -that is,
central parameters of tax systems would have to be determined and administered
by an organization equipped with supranational taxation authority. Specifically
in the European context, tax laws could be coordinated at the EU level.
The focus of supranational tax policy should be on curbing tax competition
for paper profits, since these are more mobile than real capital. Further, profit
Double tax avoidance 79
shifting is a more obvious violation of minimal fairness norms than real tax
competition.
41
Under real tax competition, a taxpayer would have to pay the
actual "price" demanded by the respective government, even though that price
may also be under harmful competitive pressure (see, for example, Sinn 1997).
This would not be the case, however, if paper profits could be shifted. In that
case, taxpayers become partial free riders: they enjoy the services offered, but
avoid paying the price for them by shifting their profits out of the country.
Although the supranationalization of tax policy sounds utopian, in Europe dis-
cussions about just such a solution are already under way. Indeed, the Treaty of
Nice has made the implementation of such measures easier. Specifically, some of
the proposals for a common consolidated corporate tax base (CCCTB) foresee,
inter alia, a solution to the problems linked to tax competition for paper profits:
42
the tax base would be defined and administered by the European Union, whereas
individual Member States would remain free to set their own respective tax rates.
Thus, a central part of national tax sovereignty would be supranationalized.
Alternative measures that could reduce tax competition for capital - paper
profits in particular - include, for instance, the introduction of an "ED CIT" -
that is, an EU-wide corporate income tax law and tax administration with the tax
and fiscal sovereignty based at ED level.43 Another measure would be the intro-
duction of a minimum level for the statutory tax rate on corporate income as
suggested in the Ruding Report (see Ruding Report 1992). But, as the current
debate on the implementation of the CCCTB shows, the political will among the
ED member countries to delegate even a part of their taxation rights to a higher
level authority is low. Thus, given the current political environment in the Euro-
pean Union, supranational measures, in order to have any realistic chance of
being adopted, will need to preserve member countries' national sovereignty in
taxation matters as far as possible. This, however, implies that tax competition
will not be curbed effectively in the near future.
Moreover, to increase the likelihood of adoption, supranational measures
should be considered that target primarily tax-induced distortions which, though
not directly related to horizontal tax competition for mobile capital, nevertheless
have side effects on it. The incentive to introduce measures of this type might be
larger for governments.
A notable example in this respect is the reduction of the distorting effects the
accrual-based classic corporate income tax system exerts, inter alia, on financing
and investment decisions of firms. Several changes in the traditional corporate
income tax system were suggested, discussed and in some cases even imple-
mented (see OECD 2007 for an overview). The focus of those measures has
been on eliminating the different tax treatment of equity and debt finance, for
instance by introducing an allowance for the remuneration of equity finance
(allowance for corporate equity, ACE) or by abolishing the deductibility of the
remuneration for debt finance (CBIT, comprehensive business income tax).
These measures result in increased efficiency in corporate income taxation
because they achieve neutrality vis-a-vis investment and/or financing decisions.
One side effect of this is that tax planning and governmental tax competition
80 M Leibrecht and T. Rixen
aiming to exploit the differential treatment of debt and equity financing (e.g. the
use of hybrid instruments) would become less effective under such systems. Pro-
viding members states with information concerning possible welfare impacts
resulting from the coordinated introduction of a more neutral corporate income
tax system by the European Commission is an important first step in this
direction.44
In any case, the sorts of far-reaching measures mentioned - in particular, a
minimum statutory tax rate on corporate income, a more neutral corporate
income tax system or an EU-wide corporate income tax law and tax administra-
tion with tax and fiscal sovereignty at EU level - must be seen at present, in light
of the current political environment, as long-term options only. In the medium
term, it is more likely that tax competition in the European Union will get fiercer
(Genschel et al., 2010).
To conclude, the given patchwork structure of DTAV rules is a good example
of how tax coordination can lead to tax-induced distortions instead of eliminating
them. This is an example of a pitfall of tax coordination. The solution to such
problems requires appropriate supranational action. Unfortunately, however,
these are currently unlikely, given present political constraints.
Appendix
AI: DTTs andforeign direct investment
Blonigen and Davies (2000)
Outcome(s): Bilateral double taxation treaties seem to exert a positive impact on
FDI stock and affiliate sales with a time lag; US inbound and outbound FDI
significantly increase with a treaty's age
Years and countries included: 1966-1992; the United States and (up to) 65
partner countries
Specification and estimator applied: Level-level specification; pooled ordinary
least square and country-pair fixed effects estimator
Dependent variable: Affiliate sales in the host country, FDI stock, FDI flows
Tax variable of main interest: Dummy variable equal to 1 if a treaty is in effect
for the country pair in a given year (treaty dummy), Length of time a treaty
has existed between the United States and a partner country in a given year
(treaty age)
Control variables: Gravity model and knowledge-capital model type variables
like real GDP of home and host country ofFDI, real GDP per capita in home
and host country of FDI, bilateral distance between countries, openness to
FDI and trade, sum of home and host countries' real GDPs, squared differ-
ence between the home and host countries' real GDP, dummy variable for the
level of a home country's skill endowment, difference in relative skilled-
labour abundance between home and host country of FDI
Double tax avoidance 81
Semi-elasticities
Table 1
Tax variable of
interest
Treaty dummy
Treaty age
Semi-elasticity
mean
27.90'
2.44
b
Semi-elasticity Semi-elasticity Standard
min. max. deviation
-29.18 132.64 55.69
-12.55 11.62 5.51
Notes
a The positive sign implies that signing a treaty increases FDI, yet