 

7' UBRARY
go; I Mich’ 931" State
University

 

 

 

This is to certify that the
dissertation entitled

THREE ESSAYS ON INTERNATIONAL TRADE AND
MULTINATIONAL FIRMS

presented by

NATHANIEL P.S. COOK

has been accepted towards fulﬁllment
of the requirements for the

PhD. degree in Economics

 

 

 

Major {Proéor’ ’5 SW

__~71m¢_3.z 100?
Date

 

MSU is an afﬁnnative-action, equal-opportunity employer

..-.-.-.-.-.-.-.-.-.-._- --- -.-..— -.-.-.-.-.- -.-.-.-i

I PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE
t; 'l U :7 a If.»
FEB I 0 2069‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/CIRC/DateDue.indd-p.1

 

THREE ESSAYS ON INTERNATIONAL TRADE AND MULTINATIONAL FIRMS
By

Nathaniel P. S. Cook

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Economics

2007

ABSTRACT
THREE ESSAYS ON INTERNATIONAL TRADE AND MULTINATIONAL FIRMS
By
Nathaniel P. S. Cook

This dissertation consists of three essays exploring the effects of economic
integration, trade policy, and tax policy on the location and production decisions of
multinational ﬁrms. In particular, it investigates the motivations for a multinational ﬁrm
to establish a foreign afﬁliate that not only makes local sales in the host country, but also
exports to foreign markets, an activity known as export platform foreign direct
investment (export platform FDI).

In recent years, exports by afﬁliates of US multinational ﬁrms have grown faster
than local sales by those afﬁliates. US multinational ﬁrms are increasingly using foreign
afﬁliates to not only serve the host country market, but also export to foreign markets.
The ﬁrst chapter of the dissertation, “Motivations for Export Platform FDI as a Strategy
for Serving Foreign Markets,” empirically investigates characteristics of countries that
tend to attract export platform F DI. A signiﬁcant contribution of this chapter is to
emphasize the importance of understanding the motivations for export platform FDI as
distinct from the motivations for F DI for local sales to the host country. The ﬁndings are
that export platform F DI by afﬁliates of US multinational ﬁrms is more prevalent in
countries have greater export market potential, as measured by the host country’s
proximity to other large markets, and are members of preferential trade agreements in

which the US is not also a member.

The second chapter of the dissertation, “Using Trade Policy to Inﬂuence Firm
Location,” examines how one country’s trade policy can affect another country’s ability
to attract FDI, and how, in the presence of economic integration in the form of a free
trade area (FTA), governments can adjust their external trade policies to inﬂuence ﬁrm
location. Interestingly, even if a country cannot itself attract FDI, its government can use
trade policy to inﬂuence whether or not other countries are able to attract export platform
FDI. A case is examined in which the formation of a F TA results in “FDI destruction,” a
multinational ﬁrm’s decision to shut down an established foreign afﬁliate. The striking
result is that this decision depends not only on the host country’s policies, but also on the
trade policies of other countries in the FTA.

The third chapter of the dissertation, “Preferential Trade Agreements and Tax
Competition with Internationally Mobile Firms,” extends existing models of international
tax competition to incorporate the activities of multinational ﬁrms headquartered outside
the competing regions. In the absence of a preferential trade agreement (PTA),
governments use positive taxes to exercise market power over the proﬁts that
internationally mobile ﬁrms must receive to locate within their borders. That is, FDI by
multinational ﬁrms is for local sales to the host country only, and foreign afﬁliates face
positive taxes in the host country. However, when two identical countries form a PTA,
tax competition drives equilibrium taxes to zero, as each country attempts to increase its
tax base within the PTA at the expense of the other country’s tax base. As a result, each
country in the PTA attracts more FDI from the rest of the world than it otherwise would,

and each afﬁliate serves as an export platform for the other country in the PTA.

ACKNOWLEDGEMENTS

Like so many things in economics, acknowledging the contributions of the many
people who have helped me along the way presents a clear identiﬁcation problem.
Nevertheless, I am certain of my indebtedness to a number of people, whom I would like
to acknowledge here.

First, I would like to thank Professor Steven J. Matusz for serving as my advisor
since my ﬁrst semester at Michigan State. I will never forget his ﬁrst words of advice
about my research, the very ﬁrst time we met, which were something to the effect of,
“First, you need to pass the Prelim. Then we can talk about your research interests.”
Right then I knew that Professor Matusz was precisely the sort of “straight-shooter” with
whom I could successfully work. Professor Matusz’s contributions have been too
numerous to list, ranging from his insights regarding the technological implications of
one of my regressors in an earlier version of the ﬁrst chapter of this dissertation to his
undeterred support in helping me work through some of the difﬁcult theoretical issues in
the third chapter. I greatly appreciate all that Professor Matusz has done for me.

Following the sage advice of Professor Matusz, I did pass the Prelim, at which
point it was time to proceed to ﬁeld courses. While I thought upon entering graduate
school that my interests were in the ﬁeld of international trade, it was Professor Jay
Wilson’s course on international trade theory that cemented my passion for the subject.
From that point forward, Professor Wilson has been an important part of my development
as an economist. His enthusiasm for this dissertation has been contagious, and his support .

throughout the entire process has been invaluable.

iv

I trace my transition from student to scholar to Dr. Susan Zhu’s course on
Advanced Topics in International Trade. It was while writing a paper for that course that
I ﬁrst encountered the idea of export platform FDI, which serves as the unifying theme in
this dissertation. Dr. Zhu’s deep knowledge of the literature in international trade is
impressive and inspiring, and her suggestions for my research have always greatly
improved it. I thank Dr. Zhu for all of her help.

A number of other people have been extremely helpful along the way. If, as
Professor Matusz suggested, the ﬁrst step to successfully completing a dissertation is
passing the Prelim, then I owe a great debt to my classmate Chris Douglas. The many
hours that we spent studying together in Marshall Hall certainly paid off. Patrice Whitely
served as a constant sounding board for both my ideas and my frustrations regarding my
research, and without that, I don’t think this dissertation would have been possible.

I would also like to thank Professor Srinivas Talluri for serving on my dissertation
committee, Professor Jeff Wooldridge for his help with my questions regarding sample
selection corrections, Professor Carl Davidson and Dr. Mike Conlin for their insightful
questions and suggestions during seminar presentations of portions of my research, and
Jennifer Carducci for her continuing support in navigating the program and its
requirements.

Finally, none of this would have been possible without the unconditional love and
support of my parents Jan and Phil Cook, my grandfather Donald Bulthaup, and my

loving wife Sarah.

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... vii
LIST OF FIGURES ......................................................................................................... viii
INTRODUCTION ........................................................................................................... 1
CHAPTER ONE: MOTIVATIONS FOR USING EXPORT PLATFORM FDI AS A
STRATEGY FOR SERVING FOREIGN MARKETS ................................................... 8
1. Introduction ..................................................................................................... 8
2. Motivations for Export Platform F DI ............................................................. 10
3. Empirical Speciﬁcation ................................................................................... 16
4. Data ................................................................................................................. 19
5. Results ............................................................................................................. 20
6. Conclusion ...................................................................................................... 25

CHAPTER TWO: USING TRADE POLICY TO INFLUENCE FIRM LOCATION... 32

1. Introduction ..................................................................................................... 32
2. The Intuition .................................................................................................... 35
3. The Model ....................................................................................................... 38
4. Equilibria ......................................................................................................... 43

Equilibria without a FTA between H and L ............................................ 44

Equilibria with a FTA between H and L ................................................ 48
5. Conclusion ....................................................................................................... 55

CHAPTER THREE: PREFERENTIAL TRADE AGREEMENTS AND TAX

COMPETITION WITH INTERNATIONALLY MOBILE FIRMS .............................. 61
1 . Introduction ..................................................................................................... 61
2. The Model ....................................................................................................... 63
3. Equilibrium in Autarky .................................................................................... 69
4. Equilibrium in a Preferential Trade Agreement .............................................. 72
5. Conclusion ....................................................................................................... 76
APPENDICES ................................................................................................................ 79
1. Deriving the representative consumer’s demand for each variety .................. 79
2. Deriving a representative ﬁrm’s proﬁt-maximizing price .............................. 80
3. Deriving the equilibrium tax in autarky .......................................................... 81
REFERENCES ............................................................................................................... 83

vi

LIST OF TABLES

1.1 Exports of Majority-Owned Afﬁliates of US Multinational Firms to Countries

Other than the US, 1999-2002 .................................................................................. 27
1.2 Summary Statistics ................................................................................................... 28
1.3 Baseline OLS Estimates ........................................................................................... 29
1.4 Robustness to Sample Selection Bias ....................................................................... 30
1.5 Total Sales, Afﬁliate Exports, and Afﬁliate Local Sales .......................................... 31

vii

LIST OF FIGURES

2.1 A Graphical Illustration of 7m .................................................................................. 57
2.2 A Graphical Illustration of 7m ................................................................................. 58
2.3 A Graphical Illustration of g .................................................................................... 59
2.4 Firm’s Equilibrium Location Decisions .................................................................... 6O

viii

INTRODUCTION

Multinational ﬁrms are the primary actors in contemporary international trade.
Recent research by Bernard et a1. (2005) shows that “roughly 90 percent of US exports
and imports...ﬂow through multinational ﬁrms” (17), and the share of trade conducted by
multinational ﬁrms is increasing through time. To understand contemporary international
trade, one must understand the activities of multinational ﬁrms.

Trade economists have long acknowledged this fact, and a large literature on the
activities of multinational ﬁrms has emerged in the ﬁeld of international trade. Initial
efforts to model the behavior of multinational ﬁrms focused on bilateral (two-country)
models. Given a ﬁrm located in one country, and consumers located in another country,
how should the ﬁrm get its product to consumers? One option is to produce in the ﬁrm’s
“home” country and export to consumers in the other country. An alternative is to
establish a foreign afﬁliate (F D1) in the “host” country where consumers are located.

This export-versus—FDI conception of the activities of multinational ﬁrms
provides many important insights into the various motivations for ﬁrms to become
multinational. Two categories of motivations for ﬁrms to become multinational are
horizontal and vertical. Markusen (1984) and Markusen and Venables (2004) model a
ﬁrm’s decision to engage in horizontal FDI as a ﬁinction of (among other factors) trade
costs and the ﬁxed costs of F D1. The higher the trade costs associated with exporting to a
foreign country, and the lower the ﬁxed costs associated with establishing a foreign
afﬁliate, the more likely a ﬁrm is to become multinational. Helpman (1984) and Helpman

and Krugman (1985) model a ﬁrm’s decision to engage in vertical FDI as a function of

factor endowments in the home and host countries and factor input requirements of
different production processes. If a ﬁrm is currently locating a production process that
intensively requires a particular factor of production in a country in which that factor is
relatively scarce (and therefore expensive), the ﬁrm may have incentive to relocate that
production process to a country in which that factor is relatively abundant (and therefore
cheaper).

While horizontal and vertical motivations for F DI provide valuable insights into
the activities of multinational ﬁrms, more recent research has explored the limitations of
bilateral models of multinational ﬁrms. In particular, treating the export-versus-FDI
decision faced by ﬁrms as a bilateral decision implicitly assumes that a ﬁrm’s decision
regarding how to serve consumers in one country is independent of the ﬁrm’s decision
regarding how to serve consumers in other countries, which is unlikely to be very
realistic.

An important activity of multinational ﬁrms that is overlooked by bilateral models
is an activity known as export platform FDI. Export platform F DI occurs when a ﬁrm
located in one country establishes a foreign afﬁliate in another country (F DI) that not
only makes sales to the host country, but also exports its products from the host country
to other countries. Data from the US Department of Commerce Bureau of Economic
Analysis (BEA) Show that in recent years more than one-quarter of the sales of majority-
owned afﬁliates of US multinational ﬁrms are exports from the host country to countries
other than the US and that the share of exports of foreign afﬁliates as a share of total sales
by afﬁliates is growing. Thus, just as it was important for trade economists to understand

the activities of multinational ﬁrms to understand international trade, it is now important

to understand the motivations for and implications of export platform FDI as an
important and increasingly prevalent activity of multinational ﬁrms. This is the
motivation for the research in this dissertation.

The ﬁrst chapter of the dissertation, “Motivations for Export Platform F DI as a
Strategy for Serving Foreign Markets,” empirically investigates characteristics of
countries that tend to attract export platform FDI. Empirical research investigating the
degree to which FDI is horizontal versus vertical has settled on a set of “standard”
characteristics that tend to be associated with FDI. Hanson et a1. (2001) ﬁnd that this
standard set of characteristics motivates sales by afﬁliates to the host country (“local
sales”) differently than it motivates exports by afﬁliates from the host country. This
suggests that the motivations for export platform FDI are different from the motivations
for F DI for local sales. Following Hanson et a1. (2001), I investigate which
characteristics of host countries tend to motivate afﬁliate exports and afﬁliate local sales
differently. Not only do I include the standard set of characteristics identiﬁed previously
in the literature, but also I include two “non-standar ” motivations for export platform
FDI: host country export market potential and host country membership in exclusive
preferential trade agreements. The hypotheses are that the closer a host country is to other
large markets, the more export platform F DI it will attract, and if exports from the host
country to certain other countries face low (or zero) tariffs as a result of the host
country’s membership in an exclusive preferential trade agreement, the more export
platform FDI it will attract. These two potential motivations for export platform FDI
would logically not appear among a set of characteristics designed to investigate the

motivations for bilateral export-versus-FDI activities of multinational ﬁrms.

I ﬁnd that both a host country’s export market potential and host country
membership in exclusive preferential trade agreements encourage afﬁliate exports more
strongly than local sales by afﬁliates. These results are robust to the inclusion of a sample
selection correction for host countries in which afﬁliate activity is (essentially) zero. This
chapter provides empirical evidence that there are important motivations for export
platform FDI that are overlooked by research that treats the activities of multinational
ﬁrms as a bilateral export-versus-FDI decision.

The result from the ﬁrst chapter that host country membership in exclusive
preferential trade agreements tends to encourage export platform FDI is examined further
in the second chapter. A paper by Raff (2004) makes the important observation that
countries may adjust their tax and trade policies in response to a decision to engage in a
preferential trade agreement. The second chapter of the dissertation, “Using Trade Policy
to Inﬂuence Firm Location” investigates a speciﬁc case of such behavior. A monopoly
ﬁrm located in one country is attempting how to get its product to consumers in two
foreign countries. The two foreign countries differ in the cost of producing the good
produced by the monopoly ﬁrm. In particular, the cost of production in one country (the
high-cost country) is sufﬁciently high that the ﬁrm will never ﬁnd it proﬁtable to locate
F D1 in that country. Thus, the ﬁrm will either export to the high-cost country from its
home country, or it will establish export platform FDI in the other (low-cost) country.
Because of the ﬁrm’s monopoly power, the high-cost country wants to levy a tariff on
imports from either the ﬁrm’s home country or the low-cost country. In the absence of a

free trade agreement between the hi gh-cost country and the low-cost country, the ﬁrm

may establish export platform F D1 in the low-cost country, and exports from the low-cost
country to the hi gh-cost country will face a positive tariff.

However, if the high cost country and the low cost country integrate in the form
of a free trade area (FTA), the high-cost country will not be able to levy a tariff on
imports from the low—cost country. Thus, the high cost country may prefer to lower its
tariff rate on imports from the ﬁrm’s home country in order to encourage the ﬁrm not to
engage in export platform FDI in the low-cost country. If the high cost country is
successful, it causes “FDI destruction.” That is, the ﬁrm would establish export platform
F D1 in the low-cost country in the absence of a FTA, but when the two countries form a
FTA, the ﬁrm ﬁnds it more proﬁtable to export to both countries than to establish export
platform FDI in the low—cost country. In this way, the high-cost country, which cannot
itself attract FDI, has considerable inﬂuence, through its trade policy, on whether or not
the low-cost country is able to attract export platform FDI.

This result illustrates once again the importance of thinking beyond bilateral
conceptions of the activities of multinational ﬁrms. FDI destruction occurs in the model
in the second chapter not because a FTA makes export platform FDI in the low-cost
country less attractive, but because in response to the FTA, a third country (the high-cost
country) adjusts its trade policy to make exporting more attractive.

The third chapter of the dissertation, “Preferential Trade Agreements and Tax
Competition with Internationally Mobile Firms” moves beyond the partial equilibrium
model in the second chapter and considers a general equilibrium model of tax
competition between two countries competing for internationally mobile ﬁrms from the

rest of the world. Standard models of international tax competition assume that countries

compete for a ﬁxed supply of capital located within the competing countries. Empirical
evidence presented by Devereux et a1. (2002) does not support the standard models of tax
competition. In particular, they suggest that countries may compete not only for capital
inside the competing regions, but also for internationally mobile capital from outside the
competing regions.

1 model two identical countries with a ﬁxed number of domestic ﬁrms that
produce varieties of a differentiated ﬁnal good and earn positive economic proﬁts.
Internationally mobile ﬁrms from the rest of the world may prefer to locate in one of the
two countries if proﬁts in those countries are higher than proﬁts available in the rest of
the world. In the absence of trade in the ﬁnal good, each country uses positive taxes to
exercise market power over the returns that internationally mobile ﬁrms require to locate
in the country. Because there is no trade in the ﬁnal good, each internationally mobile
ﬁrm that locates in one of the two countries serves only consumers in that country.

However, when the two countries form a preferential trade agreement (PTA) in
which the ﬁnal good is traded freely between the two countries, each internationally
mobile ﬁrm in each country serves as an export platform, selling not only to consumers in
the host country, but also to consumers in the other country. As a result, each country has
an incentive to lower its tax slightly below the tax in the other country, since doing so
would not affect the number of varieties of the ﬁnal good available to domestic
consumers, but by increasing its tax base, would increase the incomes that domestic
consumers have to spend on the ﬁnal good. Since both countries have the same incentive,
tax competition drives the equilibrium taxes of both countries to zero, as each country

tries to increase its tax base at the other country’s expense. Relative to the equilibrium

with no trade in the ﬁnal good, each country now attracts more internationally mobile
ﬁrms from the rest of the world, and each ﬁrm serves as an export platform.

Taken together, the three chapters of this dissertation illustrate some of the
reasons why it is important to consider export platform FDI as distinct from the
traditional exports-versus-FDI conception of the activities of multinational ﬁrms. The
ﬁrst chapter shows that export platform FDI is motivated by different factors than FDI for
local sales, in particular, the host country’s relationships (both geographical and political)
to other countries. The second chapter illustrates how the trade policies of other countries
affect a country’s ability to attract FDI. Speciﬁcally, a country that cannot itself attract
FDI may be able, through its trade policy, to affect whether or not another country is able
to attract export platform FDI. The third chapter extends traditional models of
international tax competition by introducing competition for internationally mobile ﬁrms
from outside the competing countries. In the presence of trade in the ﬁnal good between
the competing countries, each internationally mobile ﬁrm serves as an export platform,
which exacerbates tax competition between the competing regions, and generates

inefﬁciently low equilibrium taxes.

CHAPTER ONE: MOTIVATIONS FOR EXPORT PLATFORM FDI AS A

STRATEGY FOR SERVING FOREIGN MARKETS

1. Introduction

In the past two decades, a large literature on foreign direct investment (FDI) has
emerged in the ﬁeld of international trade. An important research agenda in this literature
has been to identify where multinational ﬁrms choose to locate their foreign afﬁliates and
why. Standard explanations include host country market access, tariff jumping, and factor
price differences (e. g. Brainard, 1997; Carr et al., 2001; Hanson et al., 2001; Hanson et
al., 2003; Yeaple, 2003). Hanson et a1. (2001) provide preliminary evidence that existing
research on the locational determinants of FDI has overlooked important differences in
the destination markets for afﬁliate sales. In particular, they argue that an important
expansion strategy of multinational ﬁrms is to establish a foreign afﬁliate that not only
makes local sales to the host country, but also exports its output to other countries, an
activity known as export platform FDI.

US multinational ﬁrms’ use of export platform FDI to serve foreign markets has
increased in recent years. This is evident in a dataset from the Bureau of Economic
Analysis (BEA). As shown in Table 1.1, non-US bound afﬁliate exports constitute a
substantial share of total afﬁliate sales in most industries, especially manufacturing
industries and wholesale trade. From 1999 to 2002, the value of exports by majority-

owned afﬁliates of US multinational ﬁrms to countries other than the US increased in all

13 BEA industry categories. In 1999, non-US bound afﬁliate exports accounted for more
than one-ﬁfth of total afﬁliate sales; by 2002, that number had eclipsed one-quarter.

This increasingly prevalent form of FDI merits further investigation. In this
chapter, I empirically investigate the motivations for export platform F DI as a strategy for
serving foreign markets. Hanson et. a1 (2001) showed how local sales by majority-owned
foreign afﬁliates of multinational ﬁrms and exports by those afﬁliates of are motivated
differently by the standard country and industry characteristics thought to attract F DI. My
approach differs from theirs in several important ways. First, whereas they aggregate
afﬁliate exports to the US and afﬁliate exports to countries other than the US, I focus my
attention only on non-US bound afﬁliate exports, since I am interested in export platform
FDI as a strategy for serving foreign markets. Second, whereas they look at a standard set
of country and industry characteristics thought to attract F DI, the central focus of my
analysis is to determine whether two nonstandard motivations, host country export
market potential and exclusive preferential trade agreement membership, encourage
export platform FDI. The main ﬁndings are that US export platform FDI is more
prevalent in countries that are smaller, have greater export market potential, and are
members of preferential trade agreements in which the US is not also a member.
Additionally, whereas host country preferential trade arrangement membership appears to
encourage many different types of F DI, host country export market potential uniquely
encourages export platform FDI.

This chapter is organized as follows. In section 2, I review some of the related
research on multinational ﬁrms’ decisions to invest abroad to identify the standard

motivations for FDI and then describe two potential nonstandard motivations for export

platform F DI. In sections 3 and 4, I describe my estimation approach and the data
employed in the estimation. My estimation results are reported and discussed in section 5.

Section 6 concludes.

2. Motivations for Export Platform FDI

In this section, I identify motivations for export platform F DI. First, I brieﬂy
review some of the related research on multinational ﬁrms’ decisions to invest abroad to
identify the standard motivations for FDI. Then, I describe two potential nonstandard
motivations for export platform F D1.

The existing literature on the locational determinants of FDI has attempted to
identify characteristics of host countries that tend to attract FDI. Theoretical work on this
question has produced two categories of possible motivations for ﬁrms to invest abroad:
horizontal FDI and vertical FDI.

Horizontal FDI is motivated by “market access.” When deciding how to serve a
foreign market, a multinational ﬁrm faces a choice between (among other options)
exporting to that market and establishing a foreign afﬁliate in that market (FDI). The
advantages of exporting include taking advantage of plant-level economies of scale and
avoiding the ﬁxed costs of establishing a foreign afﬁliate in the destination market. On
the other hand, by establishing a foreign afﬁliate, the ﬁrm can avoid trade costs, such as
tariffs and transportation costs. Thus, when trade costs are high relative to the magnitude

of plant-level scale economies and ﬁxed costs, ﬁrms will tend to choose FDI over

10

exports. These ideas are modeled in Markusen (1984) and Markusen and Venables
(2000)

Vertical FDI is motivated by factor price differences. If a ﬁrm is currently
locating a production process that intensively requires a particular factor of production
(e. g. labor) in a country in which that input is relatively expensive, there may be an
incentive to relocate that production process to a country in which the price of that factor
is relatively low. These ideas are modeled in Helpman (1984) and Helpman and Krugman
(1985)

Empirical research has generally concluded that most FDI is horizontal]. In the
seminal paper on the topic, Brainard (1997) ﬁnds that the share of total US sales to a
country accounted for by exports (as opposed to sales by local afﬁliates of US
multinational ﬁrms) is decreasing in trade costs and increasing in plant-level scale
economies, evidence of horizontal FDI.

Carr, Markusen, and Maskus (2001) ﬁnd that sales by afﬁliates in a host country
are increasing in host country GDP, decreasing in the squared difference between home
and host country GDP, and increasing in trade costs, all evidence of horizontal FDI.
However, they also ﬁnd that afﬁliate sales increase with the difference between the home
and host country share of skilled laborers in the labor force, evidence of vertical FDI.

Hanson et a1. (2001) suggest that vertical F DI may be more prevalent than
previous research had indicated. Their panel dataset allows them to look at changes in the
patterns of F DI across years. While they don’t run a regression to examine this particular
issue, they do note that the patterns of FDI in the 19803 and 19905 exhibit stark

differences. In the 19805, they ﬁnd a trend toward concentration of US F D1 in other

 

I For an excellent survey of the empirical literature, see Blonigen (2005).

11

OECD countries, suggesting horizontal FDI. In the 19905, however, they ﬁnd rapid
grth in the non-OECD share of afﬁliate employment, suggesting a move toward
vertical FDI. Furthermore, they ﬁnd that in their entire sample (1982-1998), growth rates
in the capital stock of manufacturing afﬁliates in OECD countries outpace growth rates in
employment, whereas in non-OECD countries, employment grth exceeds capital-stock
growth. These trends suggest the existence of both horizontal and vertical FDI.

Yeaple (2003) investigates the relative importance of horizontal and vertical
motivations for F DI. He uses cross-sectional data to regress total sales of US
multinational afﬁliates on horizontal FDI variables, such as tariff levels, F DP, and plant-
and corporate-level scale economics (as in Brainard, 1997), and vertical F DI variables,
such as a country’s average level of schooling per worker, the skill-intensity of each
industry, and the interaction between the two, establishing a “chain of comparative
advantage” (Yeaple 2003, 728). To assess the relative importance of horizontal versus
vertical motivations for FDI, after estimating the coefﬁcients for his full model, he
alternately restricts the coefﬁcients on each set of variables (horiztontal and vertical) to
be equal to zero and looks at the change in the R-squared of his regression. Using these
approach, Yeaple ﬁnds that both horizontal and vertical motivations for FDI play a role
in explaining the pattern of FDi, but that horizontal motivations seem to be relatively
more important.

Hanson et a1. (2003) examine evidence for vertical FDI in a different way. Instead
of looking at sales of afﬁliates of US multinational ﬁrms, they look at the afﬁliates’
demand for imported inputs for further processing. They ﬁnd that afﬁliate demand for

imported inputs is negatively correlated with wages for less-skilled labor (evidence of

12

vertical F DI) and positively correlated with a measure of market size (evidence of
horizontal FDI). They also ﬁnd that afﬁliate demand for imported inputs is negatively
correlated with trade costs, which emphasizes the important point that high tariffs are not
only a hurdle to he jumped by horizontal FDI, but may also be a disincentive for vertical
FDI if the production process being conducted by the foreign afﬁliate requires imported
inputs for further processing.

Two points are to be taken from this examination of existing literature on the
determinants of F DI. First, looking at these various empirical studies of F DI, a fairly
consistent set of regressors emerges. This “standard” set of country characteristics
thought to be associated with FDI includes measures of host market size (usually GDP),
average income (per capita GDP), trade costs (tariffs and/or transportation costs,
sometimes proxied for by distance between the home and host countries), factor
prices/endowments (usually proxied for by per capita GDP and/or average education),
and control variables such as corporate income tax rates and geographic measures
(distance between the home and host countries and/or an adjacency dummy variable).
While not all of these regressors are found in all of the papers discussed, most of them
will be found in any particular study. These regressors will certainly be relevant to the
present inquiry, but if we want to understand what makes export platform FDI distinct,
we need to think beyond these standard motivations for FDI.

Second, while the existing research focuses on identifying whether multinational
ﬁrms use FDI for primarily horizontal or vertical motivations, Hanson et a1. (2001 ), who
ﬁnd evidence of both horizontal and vertical motivations for F DI, write, “our ﬁndings

suggest that the literature’s benchmark distinction between horizontal and vertical foreign

l3

direct investment does not capture the range of strategies that multinationals use” (249).
In particular, it is not clear that export platform F DI ﬁts neatly into either the horizontal
or vertical classiﬁcations. Further, Hanson et. a1 ﬁnd that the “standard” motivations for
FDI identiﬁed previously in the literature have statistically signiﬁcantly different impacts
on local sales by afﬁliates to the host country, and on afﬁliate exports to countries other
than the host country (export platform F DI). That is, there is something about export
platform FDI that is different from FDI for local sales, the standard theoretical construct
in models of both horizontal and vertical FDI.

The central purpose of this chapter is to investigate two potential “nonstandard”
motivations for export platform FDI. In particular, I apply the concept of market access to
the speciﬁc case of export platform FDI. In existing studies of the determinants of FDI,
market size (GDP) is consistently found to be among the most important motivations for
F DI. By analogy, I consider in this chapter whether a measure of the export market size
of a host country is an important motivation for export platform FDI. The measure of
export market size that I use is the natural log of the inverse distance-weighted sum of
other countries’ market size, similar to the measure of market potential introduced by
Harris (1954). Harris-type measures of market potential have previously been used in the
international trade literature to explain regional productivity differences (Davis and
Weinstein, 2001). In Head and Mayer (2003), a regional Harris market potential variable
was found to be signiﬁcant and positively correlated with the location choice of Japanese
afﬁliates in the European Union. Additional evidence for the possibility that export
market potential matters for FDI is provided by Blonigen et a1. (2004), who ﬁnd that the

more FDI there is in neighboring countries, the less FDI a host country attracts, but the

14

larger the markets of neighboring countries, the more FDI a host country attracts. Baltagi
et a1. (2005) employ a similar spatial econometric approach, but ﬁnd less evidence for
export platform FDI, and more evidence for vertical and “complex vertical” FDI.

In applying the concept of market access to export platform FDI, it may not only
be geographical relationships between countries that matter. Political relationships,
embodied in preferential trade agreements, may also be relevant. Thus, I also consider in
this chapter whether host country membership in an exclusive preferential trade
agreement is a motivation for export platform FDI. Just as trade barriers may encourage
horizontal FDI (the tariff-j umping motive), a country’s membership in a preferential trade
arrangement may encourage export platform F DI by ﬁrms headquartered in non-member
countries. In a theoretical paper, Motta and Norman (1996) present a model in which
regional economic integration leads to increased export platform FDI within the region
by extra-regional ﬁrms. Neary (2002) presents a model in which the formation of an
economic union (for example, the European Union) may encourage non-union ﬁrms to
establish a single export platform within the union, but discourage multi-plant
investment. Eckholm et a1. (2003) identify conditions under which the formation of a free
trade area between a high cost country and a low cost country encourages export platform
F D1 in the low cost country. Raff (2004) also identiﬁes conditions under which economic
integration causes “FDI creation” (induces ﬁrms located outside the integrating countries
to engage in FDI when they would have served the same countries with exports in the
absence of integration). Empirically, Shatz (2004) ﬁnds that a binary variable indicating

membership in Mercosur (a preferential trade agreement in South America) is signiﬁcant

15

and positively correlated with the share of afﬁliate exports in total afﬁliate sales from

developing countries.

3. Empirical Specification

To test the hypothesis that export market potential and preferential trade
agreement membership encourage export platform FDI, I will regress a measure of export
platform activity in a country on measures of these two “nonstandard” motivations for
export platform FDI, as well as the “standard” motivations for FDI. Before describing the
full speciﬁcation of my econometric model, I will explain the particular dependent
variable employed.

An appropriate measure of export platform activity is important for my analysis.
Hanson et. a1 (2001) use the log—difference between afﬁliate exports and afﬁliate local
sales, which allows them to identify the factors that inﬂuence exports and local sales
differently. However, they do not distinguish between afﬁliate exports back to the US
and afﬁliate exports to countries other than the US. Shatz (2004) makes this distinction,
and runs separate regressions using the share of “vertical” exports back to the US in total
afﬁliate sales and the share of “horizontal” exports to countries other than the US in total
afﬁliate sales as dependent variables. He ﬁnds “striking differences” between the results
from the vertical and horizontal regressions, which suggests that the aggregation in
Hanson et a1. (2001) is not appropriate for the present inquiry. Following Hanson et a1.
(2001), but mindful of the results in Shatz (2004), the measure of export platform F DI

that I employ is EXZSU, = AEXm — ALSU, where i, j, and t index coutries, industries and

16

years, respectively. AEXU, is the natural log of non-US bound exports from country i by
majority-owned industry j afﬁliates of US multinational ﬁrms in year t in millions of US
dollars. Similarly, ALSiﬂ is the natrial log of local sales in country i by majority-owned
industry j afﬁliates of US multinational ﬁrms in year t in millions of US dollars. This
measure of export platform activity will allow me to identify the factors that inﬂuence
non-US bound afﬁliate exports and local sales differently.

The speciﬁcation of my econometric model is as follows:

EXLSU, = ,80 + [II-POTENTIAL” + ,Bz'PTAi,+ ,ByGDPi, + ,64-PCGDPI,

+ ,85-DISTANCEi + E‘s-TARIFF" + [Ry-TAX}, + ﬂg-LABORiﬁ 26 + um

The regressors POTENTIAL and PTA are the central focus of my analysis. POTENTIAL
is the Harris-type measure of export market potential mentioned previously. I constructed

POTENTIAL as follows:

 

2 GDP!“
POTENTIAL,- = .
’ k,,,~ dist”,

where GDP,“ is the purchasing power parity (PPP) value of country k’s Gross Domestic
Product (GDP) in billions of US dollars in year t and dist”, is the distance in kilometers
from country i to country k. If export platform FDI is motivated by a host country’s
proximity to large export markets, then the estimated coefﬁcient on POTENTIAL should
be positive and statistically signiﬁcant. PTA is a binary variable that takes on a value of
one if the country is a member of a preferential trade agreement in which the US is not
also a member, and takes on a value of zero otherwise. If export platform FDI is
motivated by a host country’s membership in exclusive preferential trade agreements,

then the estimated coefﬁcient on PTA should be positive and statistically signiﬁcant.

17

The next four regressors are my measures of some of the standard motivations for
FDI identiﬁed in the previous literature on the determinants of FDI location. GDP is the
natural log of the PPP value of the host country’s GDP in billions of US dollars. PCGDP
is the PPP value of the host country’s per capita GDP. DISTANCE is the natural log of
the distance in kilometers from the US to the host country. TARIFF is the unweighted
average tariff rate of the host country.

The next two regressors require a brief explanation. Previous empirical literature
looking at motivations for FDI has employed very rough measures of afﬁliate costs. For
example, because actual wage data is difﬁcult to obtain, labor costs/endowments have
oﬁen been proxied for by per capita GDP or the average education level in the host
country, which are likely to be very poor measures of the actual labor costs faced by
afﬁliates of multinational ﬁrms. In other cases, even when data are readily available, such
as statutory corporate income tax rates, these data may not accurately reﬂect the true cost
to afﬁliates, as multinational ﬁrms often take advantage of tax incentives provided by
host countries (UNCTAD 2000). For my analysis, I use measures of afﬁliate costs as
reported by the multinational ﬁrms themselves. TAX is a measure of the foreign tax
liabilities of afﬁliates; it is calculated as the value of foreign income taxes paid by
afﬁliates divided by the value of total afﬁliate sales. LABOR is a measure of the labor
costs incurred by afﬁliates; it is calculated by dividing the value of employee
compensation by the number of employees. These measures of afﬁliate costs should be
much more accurate than the measures of afﬁliate tax and labor costs used in the existing

literature.

18

Finally, Z is a vector of dummies included as controls. A full set of dummy
variables for industries and years is included in all regressions. The inclusion of the full
set of industry dummies controls for any potential time-constant industry-level ﬁxed

effects (Wooldridge 2002).

4. Data

A substantial portion of the data used in the analysis comes from the US
Department of Commerce Bureau of Economic Analysis (BEA). For the ﬁrst time in
1982 and then every ﬁfth year beginning in 1989, the BEA conducted mandatory
benchmark surveys of US direct investment abroad. In 1999, the BEA implemented two
important changes to the survey. First, they redeﬁned the industry categories they employ
in the presentation of their data. Second, they included, for the ﬁrst time, estimates for
afﬁliates that are exempt from mandatory reporting, greatly expanding the volume of
foreign direct investment activity covered by the survey. Although the benchmark
surveys are only conducted every ﬁve years, sample surveys are conducted in interim
years, yielding annual data on the direct investment activities of US multinational ﬁrms.

For my empirical analysis, I use data on FDI activity in 12 BEA industry
categories (all of the categories introduced in 1999 excluding “Utilities,” for which there
are virtually no observations of export platform FDI, and the residual “Other Industries”)
in 55 countries from 1999 to 2002. Summary statistics for my dependent and explanatory

variables are presented in Table 1.2. I restrict my attention to majority-owned afﬁliates of

19

US multinational ﬁrms. Because of the BEA’s reporting requirements, the most widely
available data are for maj ority-owned afﬁliates.

Country-level data not obtained from the BEA come from various sources. Data
on the PPP value of host country GDP and per capita GDP were obtained from the
International Monetary Fund (IMF) World Economic Outlook Database. Unweighted
average tariff rates were obtained from the World Bank. The binary variable PTA was
constructed from other sources. Finally, data on bilateral distances between countries
were obtained from the Centre d’Etudes Prospectives et d’Information Intemationales

(CEPII).

5. Results

In this section I describe the results of my empirical analysis. First, I report the
results from the ordinary least squares estimation of my baseline speciﬁcation described
in section 3. Then, I show that these results are robust to potential sample selection bias.
Finally, I investigate whether a host country’s export market potential and membership in
exclusive preferential trade agreements are motivations for FDI in general, or whether
these matter only for export platform FDI.

I ﬁrst estimated the baseline regression speciﬁcation described in section 3 by
ordinary least squares (OLS). I calculated heteroskedasticity-robust standard errors to
allow for arbitrary correlation between observations of the same country, but treating
observations of different countries as independent. Table 1.3 reports the results from this

estimation.

20

Looking at the estimation results in Table 1.3, there is strong evidence for my
hypothesis that a host country’s export market potential and membership in an exclusive
preferential trade arrangement encourage export platform FDI. The estimated coefﬁcients
of POTENTIAL and PTA are both positive and statistically signiﬁcant. Thus, US
multinational ﬁrms tend to locate export platform FDI in countries that have strong
export market potential and in countries that are members of exclusive preferential trade
agreements. Of the standard motivations for F DI, only host country market size as
measured by GDP has an estimated coefﬁcient that is statistically signiﬁcant. The
negative estimated coefﬁcient on GDP is not surprising, given the strong relationship
between local sales by foreign afﬁliates and host country market size identiﬁed
previously in the literature.

To be conﬁdent in these results, we must address a potential cause for concern.
Within the data, there are many missing or censored observations. This problem arises for
two reasons. First, the construction of my dependent variable, as described in section 3,
eliminates any observations for which either afﬁliate local sales or non-US bound
afﬁliate exports are zero. Second, the BEA suppresses certain country-industry
observations to avoid revealing conﬁdential information about the activities of a
particular multinational ﬁrm. A potential cause for concern, then, is that one or both of
these factors may be introducing sample selection bias to my baseline estimation results.

Sample selection bias has long been a concern for labor economists; more
recently, international trade economists have begun to address this problemz. The seminal
approach to addressing sample selection bias is that of Heckman (1979). In Heckman’s

two-step approach, a probit selection equation is estimated in the ﬁrst step. The results of

 

2 See, for example, Helpman et al. (2004).

21

this estimation are used to compute the inverse Mill’s ratio (IMR) for each observation,
which is then included as an additional regressor in an OLS estimation in the second step.
A valid test for sample selection bias is a test of the signiﬁcance of the estimated
coefﬁcient of the IMR, where the null hypothesis is H0: no sample selection bias
(Wooldridge, 2002).

Since there are two distinct reasons for the missing observations in my data, they
should be modeled separately. Thus, I model selection due to observations of zero
afﬁliate sales and selection due to BEA data suppression for conﬁdentiality in two
separate ﬁrst-stage probit equations and compute the IMRs from each. A valid test of
sample selection bias in this case is a test of the joint signiﬁcance of the estimated
coefﬁcients of the two IMRs.

The results of this procedure are reported in Table 1.4. The ﬁrst column of Table
1.4 reports the estimation results from the probit selection equation for observations of
zero afﬁliate sales. The second column reports the estimation results from the probit
selection equation for BEA data suppression. The third column reports the second-step
OLS estimation results, including the estimated coefﬁcients and corresponding standard
errors for the IMRs from the two ﬁrst-step probit equations. For comparison, my baseline
estimation results are reported in the fourth column. Under the null hypothesis of no
sample selection bias, the estimated coefﬁcients on the [MRS in the second-step
regression should be jointly insigniﬁcant. I fail to reject the null hypothesis of no sample
selection bias; the estimated coefﬁcients of both IMRs are individually insigniﬁcant, and

jointly insigniﬁcant (F1722 = 0.10). These results indicate that my baseline estimation

22

results are robust to any sample selection bias potentially introduced by either the
construction of my dependent variable or the BEA’s data suppression.

Thus far, I have shown that export platform FDI is motivated by both a host
country’s export market potential and a host country’s membership in exclusive
preferential trade agreements. These results are robust to potential sample selection bias.
One remaining question is whether these motivations for export platform FDI are
signiﬁcant motivations for F DI more generally, or whether these motivations are unique
to export platform F DI. To investigate this question, I separately regressed l) the natural
log of total afﬁliate sales (TAS), 2) the natural log of non-US bound afﬁliate exports
(AEX), and 3) the natural log of afﬁliate local sales (ALS) on the regresssors in my
baseline speciﬁcation described in section 3. The results from these regressions are
reported in Table 1.5.

The ﬁrst column of Table 1.5 reports the results from regressing the natural log of
total afﬁliate sales on the regressors in my baseline speciﬁcation. The fourth column of
Table 1.5 reports my baseline estimation results for comparison. These results reinforce
the point made by Hanson et al. (2001) that looking at total sales obscures important
differences in the types of F D1 in which multinational ﬁrms engage. Host country export
market potential appears to weakly discourage total afﬁliate sales, although it strongly
encourages export platform F DI.

The second and third columns of Table 1.5 report the results from regressing the
natural log of afﬁliate exports and the natural log of afﬁliate local sales on the regressors
in my baseline speciﬁcation. As described in section 3, my dependent variable is

constructed so that my baseline estimation results indicate which factors tend to motivate

23

non-US bound afﬁliate exports and afﬁliate local sales differently. The estimated
coefﬁcients from my baseline estimation results are simply the difference between the
corresponding estimates in the second and third columns. These results help clarify what
my baseline estimation results do and do not imply about the motivations for various
types of FDI. For example, the fact that PCGDP is not signiﬁcant in my baseline
estimation does not mean that it doesn’t motivate either afﬁliate exports or afﬁliate local
sales, only that it doesn’t motivate them in signiﬁcantly different ways.

Looking at my two regressors of primary interest, POTENTIAL and PTA, two
different stories emerge. In the case of POTENTIAL, the estimated coefﬁcient from my
baseline estimation results is highly signiﬁcant because whereas export market potential
encourages afﬁliate exports, it discourages local sales. In the case of PTA, the estimated
coefﬁcient from my baseline estimation results is highly signiﬁcant because although
preferential trade arrangement membership encourages both afﬁliate exports and afﬁliate
local sales, it encourages the former nearly twice as strongly as it does the latter.

These results indicate that export market potential does seem to uniquely
encourage export platform FDI, while preferential trade arrangement membership does
not. Although export market potential strongly encourages afﬁliate exports, it appears to
discourage both afﬁliate local sales and total afﬁliate sales. Thus, what previous
researchers who have observed a positive statistical relationship between market potential
and FDI have not explicitly identiﬁed is that the sole channel through which market
potential encourages F D1 is export platform F DI. Although host country membership in

exclusive preferential trade agreements encourages afﬁliate exports signiﬁcantly more

24

strongly than it encourages afﬁliate local sales, it does not uniquely encourage export

platform FDI.
6. Conclusion

Export platform F D1 is an increasingly prevalent form of FDI. Previous research
has identiﬁed a standard set of country characteristics associated with F DI, but much of
this research has aggregated all forms of FDI together, looking at data on total afﬁliate
sales, regardless of the destination of those sales. In this chapter, I have rigorously
investigated the motivations for export platform FDI as distinct from other forms of F DI.
I proceeded by identifying two potential motivations for export platform FDI that are not
part of the standard set of country characteristics thought to be associated with FDI that
has been identiﬁed previously in the literature. These two potential motivations for export
platform F D1 are a host country’s export market potential, as measured by a Harris-type
measure of the host country’s proximity to other large markets, and a host country’s
membership in an exclusive preferential trade agreement. My empirical analysis indicates
that both of these motivations are highly signiﬁcant and positively correlated with export
platform activity. Furthermore, export market potential is uniquely associated with export
platform F DI, whereas preferential trade agreement membership encourages all types of
F DI (although export platform F DI more strongly than other forms). These results
suggest that future empirical work on the motivations for FDI should not overlook the

growing importance of export platform FDI by looking at aggregate F DI measures, such

25

as total afﬁliate sales, and should pay particular attention to the motivations for export

platform FDI identiﬁed in this paper, especially export market potential.

26

Mil

Exports of Majority-Owned Affiliates of US Multinational Firms

to Countries Other than the US, 1999-2002

 

 

 

199 2 00 20 1 002
Industry Value1 Share2 Value Share Value Share Value Share
Mining 18,000 27.56% 22,041 23.85% 20,952 23.01% 25,095 26.65%
Utilities 106 0.32% N/A N/A 1,417 1.89% 1,381 3.40%
Manufacturing,Total 289,652 26.16% 332,805 28.32% 334,587 29.17% 348,331 28.82%
Food 17,209 20.27% 18,738 22.36% 19,996 23.68% 22,406 24.82%
Chemicals 55,967 29.52% 67,503 33.57% 67,376 32.89% 75,311 33.55%
”WW and 9,341 23.16% 9,372 23.82% 12,958 32.66% 13,609 33.97%
fabricated metals
Machinery 17,116 27.53% 19,465 30.28% 19,586 32.39% 24,543 40.96%
C°mp”t‘."5 and 72,117 37.09% 91,259 40.68% 84,358 41.71% 75,142 36.32%
electronic products
E'e9tr'ca' 8,714 34.39% 9,982 35.30% 9,891 36.75% 8,831 33.91%
equ1pment
T’a'Ism'tat'm 58,020 23.70% 60,935 23.94% 68,313 26.93% 73,893 27.16%
equ1pment ‘
Wholesale trade 138,441 24.87% 157,383 23.82% 189,055 28.34% 194,265 30.01%
information 8.052 11.40% 10,007 13.76% 11,003 14.30% 13,454 17.10%
.Fmance and 23,440 15.49% 25,608 13.19% 40,982 21.15% 37,811 19.03%
Insurance
Services 8,332 10.70% 9,478 11.77% 11,089 13.52% 10,405 13.12%
TOTAL, ALL , . . ,
INDUSTRIES 493,067 22.22 /. 570,022 22.734 625,762 2479/. 649,731 25.4%

1exports of majority-owned affiliates of US multinational ﬁrms to countries other than the US in

millions of US dollars

2exports of majority-owned afﬁliates of US multinational ﬁrms to countries other than the US as a
percentage of total sales of majority-owned afﬁliates of US multinational ﬁrms

Data are from the US Department of Commerce Bureau of Economic Analysis: www.bea.gov

27

TABLE 1.2

Summary Statistics

Standard
Deviation Minimum Maximum

Variable Number of
Abbreviation Observations Mean

 

 

 

Dependent
Variables
EXLS 832 -1.490 1.898 -7.290 4.581
AEX 944 5.463 2.493 0.000 11.177
ALS 1718 5.859 2.313 0.000 11.066
TAS 2138 6.234 2.254 0.000 11.355
Explanatory
Variables
POTENTIAL 2640 1.985 0.663 0.880 3.485
PTA 2640 0.627 0.484 0.000 1.000
GDP 2640 5.510 1.380 1.334 8.670
PCGDP 2640 9.401 0.856 6.709 10.971
DISTANCE 2640 8.851 0.576 6.307 9.692
TAR/FF 2616 8.565 5.872 0.000 33.000
TAX 1875 0.024 0.048 -0.375 1.000
LABOR 2135 34.457 22.388 1.379 170.000

28

TABLE 1.3

Baseline OLS Estimates

Dependent variable: EXLS

Baseline
OLS
Estimates

POTENTIAL 0.8377 ***
(0.1398)

PTA 0.5519 ***
(0.1954)

GDP -0.3571 ***
(0.1067)
PCGDP 0.1131
(0.3118)
DISTANCE 0.1539
(0.1173)
TARIFF -0.0510
(0.0325)
TAX -5.3211
(5.1442)
LABOR -0.0011
(0.0069)
constant ~2.6966
(3.8528)
Industry dummies YES
Year dummies YES
Observations 747
R-squared 0.5020

***, **, and * indicate 1, 5, and 10 percent signiﬁcance, respectively

Standard errors are reported in parentheses below corresponding
parameter estimates.

29

TABLE 1 .4

Robustness to Sample Selection Bias

 

 

Probit 1 Probit 2 OLS
Selection Selection Second-Step Baseline
Estimates Estimates Regression OLS
(zeroes) (BEA) Estimates Estimates
POTENTIAL -0.1572 " 0.1463 ‘** 0.8857 *** 0.8377 *"
(0.0830) (0.0554) (0.2860) (0.1398)
PTA 0.2485 *** -0.0010 0.5587 *** 0.5519 ***
(0.0927) (0.0670) (0.1181) (0.1954)
GDP 0.4345 *** 0.1769 *** -0.2818 -0.3571 ***
(0.0430) (0.0272) (0.3171) (0.1067)
PCGDP 0.6549 *** -0.2029 *" 0.0765 0.1131
(0.1041) (0.0799) (0.4294) (0.3118)
DISTANCE -0.0883 -0.1091 ** 0.1169 0.1539
(0.0908) (0.0536) (0.2079) (0.1173)
TARIFF 0.0194 * -0.0420 *** -0.0643 -0.0510
(0.0111) (0.0086) (0.0814) (0.0325)
TAX 4.0252 *“ -1.0278 -5.5471 -5.3211
(1.4125) (1.1367) (3.5660) (5.1442)
LABOR 0.0111 *‘* -0.0006 -0.0013 -0.0011
(0.0030) (0.0021) (0.0040) (0.0069)
constant -7.7815 "" 1.8403 "' 0.1826 -2.6966
(1.3209) (0.9540) (0.5471) (3.8528)
Inverse of Mill's Ratio 0.6153
from Probit 1 (zeroes) (3.2085)
Inverse of Mill's Ratio -3.0484
from Probit 2 (BEA) (2.5583)
Industry dummies YES YES YES YES
Year dummies YES YES YES YES
Total Observations 1717 1717 1717
Censored Observations 970
Uncensored Observations 747 747

***, **, and * indicate 1, 5, and 10 percent signiﬁcance, respectively

 

Standard errors are reported in parentheses below corresponding parameter estimates.

30

TAB_L_E 1.5

Total Sales, Affiliate Exports, and Afﬁliate Local Sales

POTENTIAL

PTA

GDP

PCGDP

DISTANCE

TA RIFF

TAX

LABOR

constant

Industry dummies

Year dummies

Observations
R-squared

***, **, and * indicate 1, 5, and 10 percent signiﬁcance, respectively

 

Dependent Variable
TAS AEX ALS
-0.2718 * 0.3952 * -0.4425 ***
(0.1503) (0.2216) (0.1421)
0.7899 *** 1.2547 *** 0.7028 ***
(0.1782) (0.2788) (0.1673)
0.6324 *** 0.4417 *** 0.7988 ***
(0.0683) (0.1254) (0.0629)
0.6124 *** 0.7614 * 0.6482 ***
(0.2287) (0.4086) (0.1905)
-0.5113 **' -0.3418 * -0.4957 ***
(0.1030) (0.1765) (0.0941)
-0.0260 -0.0697 -0.0187
(0.0266) (0.0454) (0.0213)
-8.5334 -11.4559 -6.1348
(5.5170) (8.8835) (5.0209)
0.0128 *** 0.0124 0.0135 ***
(0.0034) (0.0082) (0.0033)
2.6411 -2.1358 0.5608
(2.7331) (5.0564) (2.3780)
YES YES YES
YES YES YES
747 747 747
0.6673 0.5409 0.7381

 

Baseline
OLS

Estimates

0.8377
(0.1398)

0.5519 ***
(0.1954)

-o.3571
(0.1067)

0.1131
(0.3118)

0.1539
(0.1173)

-0.0510
(0.0325)

-5.3211
(5.1442)

-0.0011
(0.0069)

-2.6966
(3.8528)

YES
YES

747
0.5020

Standard errors are reported in parentheses below corresponding parameter estimates.

31

CHAPTER TWO: USING TRADE POLICY TO INFLUENCE FIRM LOCATION

1. Introduction

As multinational ﬁrms have emerged as the dominant players in the global
economy, trade economists have tried to understand the causes and consequences of F DI.
An important research agenda has been to identify country characteristics that inﬂuence a
multinational ﬁrm’s decision to serve a country via exports or FDI. Two important
ﬁndings of this research are relevant to the current investigation. First, the “tariff-
jumping” motive for FDI has been well-documented (Caves, 1982; Brander and Spencer
1987; Motta, 1992; Hwang and Mai, 2002). Put simply, the higher a country’s tariffs on
imports, the more likely a multinational ﬁrm will be to establish a foreign afﬁliate within
that country, thereby “jumping” the high tariff barrier. Thus, high (low) tariffs will tend
to encourage (discourage) FDI by multinational ﬁrms. Second, economic integration
tends to encourage multinational ﬁrms headquartered outside a preferential trade
agreement to establish a foreign afﬁliate inside the agreement to take advantage of lower
internal barriers to trade (Motta and Norman, 1996; Neary, 2002; Ekholm et al., 2004).
This is a logical extension of the tariff-jumping motive for F DI: If two countries form a
free trade area, it is possible to bypass one country’s high tariff barriers by establishing a
foreign afﬁliate in the other country and exporting inside the free trade area. The
important lesson from this research is that a country’s trade policy and membership in
preferential trade agreements inﬂuence multinational ﬁrms’ decisions regarding the

location of FDI.

32

As a rich tradition in public economics tells us, trade policy is not the only type of
government policy that inﬂuences ﬁrms’ location decisions. An extensive literature on
tax competition for internationally mobile capital suggests that governments also use tax
policy to inﬂuence the location of FDI. While such international tax competition is often
wasteﬁrl, and results in inefﬁciently low tax rates (Wilson, 1999), J aneba (1998) shows
that in the presence of imperfect competition, tax competition for internationally mobile
ﬁrms can actually be welfare-improving.

Although many papers have explored the effects of either tariffs or taxes on the
location of investment, very little research has investigated the effects of both tax and
trade policy on ﬁrm location. Raff (2004) considers both tax and trade policy by
examining how a hi gh-cost country and a low-cost country will compete in taxes and
tariffs to inﬂuence the location of FDI by a foreign monopolist, and how each country
will adjust its policies in response to economic integration in the form of a free trade
agreement (FTA). Raff makes several claims about how governments will respond to
integration, and thus how integration will ultimately affect the ﬁrrn’s location choice.
First, Raff claims that a FTA may cause “F DI creation,” meaning that in the absence of a
FTA, governments choose policies to encourage the ﬁrm to export rather than invest, but
in the presence of a FTA, governments adjust their policies to induce the ﬁrm to establish
a foreign afﬁliate in the low-cost country. Second, Raff claims that if, in the absence of a
F TA, the ﬁrm invested in both countries, the FTA will cause “FDI consolidation,”
meaning that the ﬁrm will shut down its afﬁliate in the high-cost country and serve both
countries from its afﬁliate in the low-cost country. Finally, Raff claims that a FTA cannot

cause “FDI destruction,” meaning that the ﬁrm operates at least one foreign afﬁliate in

_33

the absence of a F TA, but in the presence of a F TA, the ﬁrm will serve both countries
with exports.

Using Raff’s basic model, this paper shows that the high-cost country has
considerable power to inﬂuence a multinational ﬁrm’s investment location decision using
only trade policy, even when it cannot itself attract FDI. Speciﬁcally, the high-cost
country may be able to prevent the ﬁrm from investing in the low-cost country even
when the low-cost country wants to attract FDI. Furthermore, in contrast to Raff, this
paper shows that a FTA may cause F DI destruction. In other words, the ﬁrm may choose
to invest in the low-cost country in the absence of a FTA, but when the high-cost country
and the low-cost country form a FTA, the hi gh-cost country is able to use its trade policy
to convince the ﬁrm to shut down its foreign afﬁliate in the low-cost country.

How can two papers using the same model generate different results? This paper
considers a case of the model not considered by Raff. Whereas Raff assumes that the ﬁrm
could proﬁtably serve both countries with an afﬁliate in the high-cost country, this paper
assumes that the ﬁrm could never proﬁtably invest in the high-cost country. In other
words, Raff assumes that the cost of production in the high-cost country is “not too much
bigger” than the cost of production in the low-cost country, but this paper assumes that
the cost of production in the high-cost country is “very high.” As a result, the high-cost
country will not be able to use its tax rate to inﬂuence the ﬁrm’s location choice.
However, since the government of the high-cost country cares whether or not the ﬁrm
invests in the low-cost country, it will use its trade policy to inﬂuence the ﬁrm’s location

decision.

34

While the different results generated by this paper and Raff (2004) are primarily
due to the different cases of the model being considered, the analysis in this paper sheds
light on a mistake in Raff (2004). In the model, a FTA is unambiguously welfare-
reducing for the high-cost country if it imports from the low-cost country, since a FTA
eliminates the potential for positive tariff revenue. Raff (2004) mistakenly claims that a
F TA between the two countries can create FDI that is welfare-improving for both
countries’. The reason that FDI destruction occurs in the case considered in this paper is
precisely because the hi gh-cost country actively uses its trade policy to prevent the low-
cost country from attracting FDI. Both because he mistakenly claims that FDI creation
could be welfare-improving for the hi gh-cost country in his model, and because he does
not consider the speciﬁc case of the model analyzed in this paper, Raff (2004) mistakenly
claims that a FTA cannot cause F DI destruction. This paper shows when and why F DI
destruction may occur.

The rest of the paper is organized as follows. Section 2 provides the intuition for
the results to be derived in the rest of the paper. Section 3 outlines the speciﬁcs of the
theoretical model. Section 4 analyzes equilibria both without and with a FTA between H

and L. Section 5 concludes.
2. The Intuition
In the model, a monopoly ﬁrm chooses how to supply two foreign countries. The

ﬁrm may choose to export to one or both countries from its existing domestic production

facility, or it may choose to establish a foreign afﬁliate (FDI) in one or both of the foreign

 

3 FDI creation in Raff (2004) cannot be welfare-improving for the high-cost country (Cook 2006).

35

 

countries. If the ﬁrm chooses to establish a foreign afﬁliate, it must decide whether to use
that afﬁliate for local sales in the host country only, or to also use the afﬁliate as an
export platform for sales to the other foreign country.

The two foreign countries differ in the marginal cost of producing the monopoly
ﬁrm’s good, so we can refer to them as the high-cost country and the low-cost country.
Both countries care about the ﬁrm’s chosen supply strategy, and can use tax and/or trade
policy to inﬂuence the ﬁrm’s decision. If the ﬁrm chooses to supply a country with
exports, that country’s optimal tariff is positive, because of the ﬁrm’s monopoly power
(Vousden 1990). If the ﬁrm establishes a foreign afﬁliate in a country, that country wants
to extract the maximum possible tax revenue from the foreign afﬁliate. For this reason,
attracting a foreign afﬁliate that not only makes local sales in the host country market, but
also exports to the other country (export platform FDI) may be particularly beneﬁcial,
since export platform F DI represents a much larger tax base than FDI for local sales only.

The cost difference between the two countries affects which policies are relevant
for each country. It is assumed that the marginal cost of production in the hi gh—cost
country is sufﬁciently high that the hi gh-cost country can never attract FDI. Thus, the
hi gh-cost country’s tax policy is irrelevant, since it will never have a foreign afﬁliate to
tax. Similarly, the low-cost country’s tariff on imports from the high-cost country is
irrelevant, since the ﬁrm will never produce in the high-cost country. This leaves the
high-cost country with only trade policies: a tariff on imports from the ﬁrm’s home
country, and a tariff on imports from the low-cost country. The low-cost country can use
both its tax and its tariff on imports from the ﬁrm’s home country to inﬂuence the ﬁrm’s

decision.

36

The surprising result is that despite its own inability to attract the ﬁrm, the high-
cost country, through its trade policy, can exert considerable inﬂuence on the ﬁrm’s
decision to establish FDI in the low-cost country. In particular, although the low-cost
country always (weakly) prefers attracting export platform FDI to attracting FDI for local
sales only, in the absence of a F TA between the hi gh-cost country and the low-cost
country, the low-cost country is only able to attract export platform F DI when the high-
cost country prefers importing from the low-cost country to importing from the ﬁrm’s
home country. In a sense, the high-cost country is able to impose its preferences on the
low-cost country in the absence of a FTA.

Perhaps even more surprising is the inﬂuence that the high-cost country has when
the hi gh-cost country and the low cost country form a FTA. Note that a FTA eliminates
one of the hi gh-cost country’s two policy tools, since a FTA restricts the high-cost
country’s tariff on imports from the low-cost country to be zero. Left with only its tariff
on imports from the ﬁrm’s home country, the high-cost country can still prevent the ﬁrm
from establishing export platform FDI in the low-cost country. In some cases, the high-
cost country can lower its tariff on imports from the ﬁrm’s home country enough to
convince the ﬁrm to shut down an existing foreign afﬁliate in the low-cost country. To
use Raff’s term, a F TA can cause FDI destruction. The intuition for why this could
happen is provided below.

The high-cost country’s only decision is where it would prefer to import from.
Assuming it is free to set whatever tariff it chooses on imports from either the ﬁrm’s
home country or the low-cost country, the high-cost country prefers to import from

whichever location. has the lower cost of production. If the cost of production in the low-

37

cost country is lower than the cost of production in the ﬁrm’s home country, then the
high-cost country prefers to import from the low-cost country. In the absence of a FTA,
the ﬁrm may establish export platform FDI in the low-cost country, and if necessary, the
high-cost country may be willing to lower its tariff on imports from the low-cost country
in order to assist the low-cost country in attracting export platform FDI. However, a FTA
between the low-cost country and the high-cost country eliminates tariff revenue for the
high-cost country on imports from the low-cost country. If the low-cost country’s cost of
production advantage over the ﬁrm’s home country is not sufﬁciently large to
compensate for this loss of tariff revenue, then the hi gh-cost country will attempt to lure
the ﬁrm out of the low-cost country by lowering its tariff on imports from the ﬁrm’s
home country. If the hi gh-cost country can convince the ﬁrm to source its exports to the
high-cost country from the ﬁrm’s home country rather than from the low-cost country,
then the low-cost country will no longer be able to attract export platform FDI, and the

FTA will cause FDI destruction.

3. The Model

A monopoly producer of good X operates a plant in country F. Consumers of
good X reside in two foreign countries, H and L. Demand for X in country i e {H, L}, x,,
is given by:

x,- = 1 —p.- (1)

where p, is the price the monopoly ﬁrm charges in country i.

38

The ﬁrm must choose the proﬁt maximizing supply mode for serving consumers
in H and L. One option is to export to both H and L from the established plant in F, where
the constant marginal cost of production is f < 1. However, the ﬁrm may also choose to
establish a foreign afﬁliate in either H or L or both, and pay a ﬁxed (but not sunk) cost of
G per foreign afﬁliate. If the ﬁrm invests abroad, the constant marginal costs of
production in H and L are h and 1, respectively, with l< h < 1. In other words, H is the
“high-cost” foreign location, and L is the “low-cost” foreign location.

The ﬁrm has six potential supply strategies:

1) Export to both H and L from F.

2) Export to H, and establish a foreign afﬁliate for local sales in L.

3) Export to L, and establish a foreign afﬁliate for local sales in H.

4) Establish a foreign afﬁliate in L for both local sales in L and export sales to H.

5) Establish a foreign afﬁliate in H for both local sales in H and export sales to L.

6) Establish a foreign afﬁliate in both L and H, using each for local sales only.

The governments of H and L are not indifferent about the ﬁrm’s decision, and have
policy tools at their disposal to inﬂuence the ﬁrm’s supply strategy. Speciﬁcally, H and L

each have three policy instruments:

1) a tariff rate on imports from F, r”? i e {H, L}
2) a tariff rate on imports from the other country, rjj i, j e {H, L}, 1' ¢ j
3) a tax rate on proﬁts from ﬁrm sales originating in i, t,- z' e {H, L}

The governments of H and L choose available policy tools noncooperatively to maximize
their respective social welfare functions, which consist of consumer surplus and tax or

tariff revenue.

39

The central question of this chapter, then, is how the governments of H and L will
use their available policy tools to inﬂuence the ﬁrm’s supply strategy. Furthermore, how
will these decisions change in the presence of a free trade agreement (FTA) between H
and L? Note that a F TA between H and L restricts each govemment’s available policy
tools. In particular, a F TA restricts rm, = rer = 0.

The timing of the interactions between H, L, and the ﬁrm can be summarized as follows:

Stage 1: H and L choose available policy instruments noncooperatively.
Stage 2: The ﬁrm observes the policies of H and L and chooses its supply strategy.
Stage 3: The ﬁrm chooses pH and pL to maximize proﬁts, given its supply strategy.

The (sub game perfect) equilibria of the model, both with and without a FTA between H
and L, can be solved for using backward induction.
We begin by looking at Stage 3, in which the ﬁrm chooses its proﬁt maximizing prices,
taking as given the values of the taxes and tariffs chosen by H and L, and its own chosen
supply strategy.

For illustration, consider that the ﬁrm’s chosen supply strategy is 4), in which the
ﬁrm establishes a foreign afﬁliate in L to serve customers in both L and H (export

platform FDI). The ﬁrm solves the following proﬁt maximization problem:

gnax (1 — IL)[(pL - l)(1 —pL) + (PH — l — I‘HLXI -pH) — G] (2)
H aPL

The ﬁrst two terms inside the square brackets represent the ﬁrm’s pre-tax proﬁts from
serving consumers in L with locally-produced X (gross of the ﬁxed cost of establishing a
foreign afﬁliate in L). The second two terms inside the square brackets represent the
ﬁrm’s pre-tax proﬁts from serving consumers in H with exports from L, facing H’s tariff

rate against imports from L, rm. The ﬁnal term inside the brackets is the ﬁxed cost of

40

establishing a foreign afﬁliate in L. Because all of these sales originate in L, all of the
proﬁts (net of the ﬁxed cost of establishing a foreign afﬁliate) are subject to L’s proﬁt tax
rate, Q.

The proﬁt maximizing prices are pL = (1 + l)/2 and pH = (1 + l + rHL)/2. The
corresponding quantities are xL = (1 — l)/2 and Jr” = (1 — l — rHL)/2. This generates an after-

tax proﬁt for the ﬁrm of

(1 _,L)[(1 —41)- +04:...) -0] (3)

 

 

Given these prices, the social welfare in L is

(1—1)2+t (1—1)’+(1—l-rm)’_G (4)
8 " 4 4

 

 

The ﬁrst term in (4) is the consumer surplus from consuming locally-produced X. The

second term in (4) is L’s tax revenue. The social welfare in H is

(5)

 

(1_l_rHL)2 (l—l—rllL)
+r ——
8 ”L 2

The ﬁrst term in (5) is the consumer surplus from consuming X imported from L. The
second term in (5) is H’s tariff revenue.

H and L will choose their available policy instruments to maximize social
welfare. Note that (4) is strictly increasing in tr, so assuming that L wants to attract FDI,
it will want to set tL as high as possible as long as the ﬁrm will still invest. The same is
not true of (5). Assuming that H’s choice of rHL will not affect the ﬁrm’s supply strategy,

H faces the following maximization problem:

 

maX(1—1_rHL)2 +rHL(1—l—rnr) (6)
2

rHL 8

41

H’s “optimal tariff” on imports from L is thus rHL* = (1 —— l)/3. Given this tariff rate, H’s
social welfare from importing from L is (1 — [)2/6, and the ﬁrm’s pre-tax proﬁt from
exporting from L to F (gross of the ﬁxed cost of investing in L) is (1 — [)2/9.

To this point, the model is very similar to the model in Raff (2004). The major
difference between the present model and Raff’s model is a different assumption about
the cost parameters in the model. So far, the model assumes that I < h < 1. However, Raff
further assumes that the cost of production in H (h) is “not too much bigger” than the cost

of production in L (1)4. In contrast, this paper makes the following assumption:

 

_ 2
Assumption 1: G > ((1 2h) J

Assumption 1 does not restrict the difference between h and l, but does imply that h is
large. In particular, h is large enough that under no circumstances will the ﬁrm choose to
invest in H. This assumption has important consequences for the model.

First, it restricts the number of reasonable supply strategies for the ﬁrm from six
to three. In particular, since the ﬁrm will never choose to invest in H, the ﬁrm’s
remaining reasonable supply strategies are to export to both H and L from F, to export
from F to H and establish a foreign afﬁliate in L for local sales in L only, or to establish a
foreign afﬁliate in L for both local sales in L and export sales to H.

Second, it restricts the relevant policy tools for both H and L. In particular, since
H will never attract FDI, tH is irrelevant. By the same token, since H will never attract

FDI, L will never import from H, and thus rLH is irrelevant. This leaves each country with

 

 

5(1—h)2 -(l —l)2
8

4 Using the notation in this paper, Raff (2004) assumes G < . For both this inequality

and Assumption I to be satisﬁed requires 1 > h. Thus, this paper considers a case not considered in Raff
(2004).

42

two relevant policy tools in the absence of a FTA. L is left with a tariff rate against
imports from F, m, and a tax rate on proﬁts generated from sales originating in L, tL. It is
signiﬁcant to note that H is left with only trade policy tools: a tariff rate on imports from
F, rm, and a tariff rate on imports from L, rHL. Thus, even though H cannot itself attract
FDI, H cares from which country it imports, and may be able to use its trade policy tools
to inﬂuence the ﬁrm’s decision. Note also that if H and L form a FTA, L does not lose
any policy tools (since it never imports from H anyway), but H loses one of its two policy
tools, since a FTA restricts rm = 0. The implications of this will be explored in greater
detail below. The following section derives the equilibrium policies of H and L and the

equilibrium supply strategy of the ﬁrm, both without and with a F TA between H and L.
4. Equilibria

As explained in Section 2, an important consideration is the relationship between I
and f. For illustration, we will consider the case in which I < f 5 . This means that the low-
cost country has a cost of production advantage over the ﬁrm’s home country. Further,

we will assume:

Assumption 2: [ﬂ] Big—21):]
6 8

 

5 Results for the case in which I > fare qualitatively similar to the results derived from the case considered
here. The central difference is that in the absence of a FTA between H and L when I > ﬂ L will not attract
export platform FDI. If L attracts FDI, it is for local sales only. A FTA may still cause FDI destruction in
this case, and for the same reasons: H lowers its tariff rate against imports from F to prevent the ﬁrm from
locating FDI in L.

43

Assumption 2 implies that although the low-cost country has a cost of production
advantage over the ﬁrm’s home country, this cost of production advantage is smallb.
Assumption 2 has two important interpretations. From H’s perspective, Assumption 2
means that despite L’s cost of production advantage over F, H prefers to import from F
and levy its optimal tariff than to import from L tariff-free. From L’s perspective,
Assumption 2 means that the consumer surplus that consumers in L enjoy from
consuming locally-produced X is smaller than the social welfare in L from importing X
from F and levying L’s optimal tariff. This means that for L to prefer attracting FDI to

importing from F, L must be able to collect some positive amount of tax revenue.

Equilibria without a F T A between H and L

First, consider equilibria in the absence of a FTA between H and L. Since I <ﬁ
there is an ex ante cost advantage to the ﬁrm from locating production in L. However,
depending on both the exogenous ﬁxed cost of establishing a foreign afﬁliate in L (G)
and the endogenous policy choices of H and L, the ﬁrm may choose to locate production
in F, despite the higher marginal cost of production. If the ﬁrm locates production in L, it
has a second decision about whether to export to H from F or from L.

H has two available trade policies to inﬂuence the ﬁrm’s location decision: rm
and rHF. If the ﬁrm chooses to establish export platform FDI in L and export from L to H,

H’s social welfare will be maximized by rHL’I‘ = (l - l)/3. Similarly, if the ﬁrm chooses to

 

6 Even if Assumption 2 is violated and L has a large cost of production advantage, 3 FTA may cause FDI
destruction. If Assumption 2 is violated, then in the absence of a FTA H prefers to subsidize imports from L
rather than import from F. If a FTA between H and L prevents H from subsidizing imports from L, a FTA
may cause FDI destruction.

44

locate production in F and export from F to H, H’s social welfare will be maximized by
rm?“ = (l —f)/3.Since we are looking at the case in which I < f, H prefers importing from
L at rHL’“ and earning a social welfare of (1 — [)2/6, to importing from F at rHF* and
earning a social welfare of (l —f)2/6. In fact, because of L’s cost of production advantage,
H would prefer importing from L and levying rm < rHL“ to importing from F and levying

rHF*. L will prefer to import from L as long as:

{(1 —[ -8-rHI.)- ]+ rHL[g;_l_2:ﬁﬂ]Z[(_l_—6L)_] (7)

The left-hand side of (7) is L’s social welfare from importing from L and levying rm.

 

The right-hand side of (7) is L’s social welfare from importing from F and levying rHF*.
By Assumption 2, (7) is violated for rm = 0, but since I < f, (7) is satisﬁed for rHL = rHL“.
Deﬁne 7m 6 [0, rHL*] as the value of rHL that satisﬁes (7) with equality. TM is the
smallest value of rm that H would be willing to set to encourage the ﬁrm to establish
export platform FDI in L. Note that the left-hand side of (7) is strictly increasing in rm in
( 7m , rm“), meaning that although L still prefers importing from L and levying rm

6 ("Fm , rHL’I‘) to importing from F and levying rHF = rHF’“, L’s social welfare from
importing from L increases the higher tariff rate it sets against imports from L (up to
rHL’I‘). H’s social welfare from importing from L and from F, as well as 7”,, rHF’“, and

rHL’f, are illustrated graphically in Figure 2.1.
Since H prefers importing from L to importing from F as long as (7) is satisﬁed,

and since (7) can be satisﬁed for for rm 6 (7m , rHL*), H can raise its tariff rate against

imports from F to a level that would eliminate any proﬁts for the ﬁrm from exporting

from F to H, rHF = l — f. This is effectively a “stick-and-carrot” strategy; because L

45

prefers importing from L to importing from F, it can raise its tariff rate against imports
from F (the stick) to discourage the ﬁrm from locating production in F, and if necessary,

L may lower its tariff rate against imports from L as low as 7}“ (the carrot) to encourage

the ﬁrm to export to H from L.

Thus, if the ﬁrm locates production in L, it will use that investment as an export
platform for H, since H can employ the stick-and—carrot strategy described above to
induce the ﬁrm to export to H from L rather than from F. The ﬁrm’s decision then, is
between establishing export platform FDI in L and exporting to both H and L from F. For

the ﬁrm to prefer establishing export platform FDI in L requires:

(1_tL)[(1—l)’+(1-l-rrri) 4;} (l—I—rs) (8)

 

 

 

4 4 _ 4
The left-hand side of (8) is the ﬁrm’s after-tax proﬁts from establishing export platform
FDI in L, where exports from L to H would face a tariff of rm. The right-hand side of (8)
is the ﬁrm’s proﬁt from exporting to both H and L from F, where exports to H would face
mp = l — f, and exports to L would face rLF. Whether or not (8) is satisﬁed depends on
both the exogenous ﬁxed cost of investment (G) and L’s endogenously chosen policies
(tL and rur). In particular, note that if L prefers to attract export platform F DI, then like H,
L can discourage imports from F by setting rip = l — f, and (8) will be satisﬁed with
equality for tr = 1. On the other hand, if L prefers to import from F, it can set Q = 1 and
rLF < l — f and (8) will be violated. In neither case does H’s choice of rHL matter for the

ﬁrm’s decision (although it does inﬂuence L’s preference).

46

If L imports from F, L’s social welfare is maximized by levying rLF’I‘ = (l —f)/3.
L’s social welfare from attracting export platform FDI is complicated, because it depends
not only on the tax rate chosen by L, Q, but also on the exogenous ﬁxed cost of
investment, G, and on the tariff rate levied by H against imports from L, rHL. For L to

prefer attracting export platform FDI to importing from F requires:

(1-1)2 +1 (1-1)2 +(1-1—rm)2 _G >9;f_)2 (9)
8 L 4 4 — 6

 

The left-hand side of (9) is L’s social welfare from attracting export platform FDI, where

exports from L to H face a tariff of rHL. The right-hand side of (9) is L’s social welfare
from importing from F and levying rLF*. Deﬁne Pm as the value of rm that satisﬁes (9)
with equality for tL = 1. From (7), the smallest value of rHL that H would be willing to set
to encourage the ﬁrm to invest in L is 7”,. Thus, the largest value of G for which (9)

could be satisﬁed is

(10)

 

 

Q(7HL)=[3(1;1)2+0427...) 31—6102]

If G > G (7m ), then f”, < FHL and H will not be willing to set rHL low enough for L to

prefer attracting export platform FDI to importing from F. If L prefers importing from F

to attracting export platform FDI, it will set tL = 1 and UP = rm“ and the ﬁrm will choose

to locate production in F, since (8) will be violated. If G SQ (7m ), H will set rm =
PM and (9) will be satisﬁed with equality, so L will set t1, = 1 and rip = l — f and the ﬁrm

will locate export platform F D1 in L. If G SQ ( rHL*) where

(11)

 

350—1): 31-02]

Gr *=
_(IIL)[ 72 6

47

then H will set rm = rHL’“ and (9) will not be binding, 50 L will strictly prefer to attract

export platform FDI.

Proposition 1 . In the absence of a FTA between H and L:

(a) ifG SQ (rHL*), H will set (rHF = 1 —f, rHL = rHL*), L will set (ru: = 1 —f, tL = l), and
the ﬁrm will establish export platform FDI in L in equilibrium.

(b) ifGe (Q (rHL*),Q (7m )], H will set (rHF = l —f, rm = 7"”, ), L will set (er = 1 —f, tL =
1), and the ﬁrm will establish export platform F D1 in L in equilibrium.

(0) ifG >Q (7m ), H will set (rt/F = rHF*, rm =77”, ), L will set (er = up“, tL = 1), and the

ﬁrm will export to both H and L from F.
Equilibria with a F T A between H and L

Now consider equilibria in the presence of a F TA between H and L. A F TA
restricts rHL =0. Given Assumption 2, H prefers importing from F and levying er* to
importing from L tariff-free. Even though H’s social welfare from importing from. F is
maximized by er’“, H may be willing to set rHF < mp“ to encourage the ﬁrm to export
from F to H rather than from L to H. For H to prefer importing from F and setting rm to

importing tariff-free from L requires

[(1_f;rHF) ]+rHF[(1"f2—FHF)] Z[(_1—8i) (12)

 

 

The left-hand side of (12) is H’s social welfare from importing from F and setting rHF.

The right-hand side of (l 2) is H’s social welfare from importing tariff-free from L. Since

48

l < f, ( 12) will be violated for rHF = 0, but by Assumption 2, (12) will be satisﬁed for er“.
Deﬁne FHF e [0, rHF*] as the value of rHF that satisﬁes (12) with equality. Thus, 7m is the
smallest value of rim that H would be willing to set to prevent the ﬁrm from investing in
L. A graphical illustration of 7”,. is provided in Figure 2.2.

Engaging in export platform FDI in L is much more attractive to the ﬁrm in the
presence of a FTA than in the absence of a FTA. Since rm = 0, the ﬁrm’s pre-tax proﬁts
from export platform FDI in L are larger than they would be in the absence of a FTA
between H and L (when H would set rHL > 0).

Since any proﬁt that the ﬁrm generates in L is L’s tax base, L will also be more
willing to attract FDI in the presence of a FTA than in the absence of a FTA. For L to

prefer attracting export platform FDI to importing from F requires

[(1—1)’]+Q[(1-1)2+(1—l)’_G] Z[(1—f)’] (13)

 

8 4 4 6

The ﬁrst term on the left-hand side of (13) is L’s consumer surplus from consuming
locally-produced X. The second term on the left-hand side of (13) is L’s tax revenue from
the ﬁrm’s proﬁts from local sales in L and tariff-free exports from L to H, net of the ﬁxed
cost of establishing the afﬁliate in L. The right-hand side of (13) is L’s social welfare
from importing from F and levying rLF*.

Whether or not (13) is satisﬁed depends not only on the parameters 1, f and G, but
also on L’s endogenously chosen tax rate, tr. Assuming for a moment that L could
impose its maximal tax rate, tL =1, the largest value of G for which (13) could be satisﬁed

is

49

Qll

 

=(15(1—1)’—4(1—f)2] (14)

24

That is, G is the largest possible ﬁxed cost for which L would be willing to attract export
platform FDI. IfG <G , there are values oftL for which (13) could be satisﬁed. IfG >G ,

L will prefer importing from F to attracting export platform FDI. Note that G >Q ( 7m ),

conﬁrming that L is more willing to attract FDI in the presence of a FTA than in the

absence of a FTA.
In what follows, assume that G < G . Deﬁne t L (G) as the tax rate that satisﬁes
(13) with equality for a given value of G. For any tL < t 1. (G), (13) will be violated, and L

will prefer importing from F and levying er’“ to attracting export platform FDI. For any

tL > t_L (G), (13) will be satisﬁed, and L will prefer attracting export platform FDI to
importing from F. Thus, t L (G) is the lowest possible tax rate that L would be willing to

set to attract export platform FDI, given the ﬁxed cost of investment faced by the ﬁrm.

6%

Note that > 0. In other words, the larger the ﬁxed cost of investment, the smaller

 

the ﬁrm’s (pre-tax) proﬁt from export platform F D1 in L, and thus the larger that tax rate
that will be necessary for L to be willing to attract export platform FDI.

The preceding analysis identiﬁes when L is willing to attract export platform FDI,
but for L to be able to attract export platform FDI, the ﬁrm must prefer establishing
export platform FDI in L to exporting to H and L from F7. Note that the ﬁrm’s decision to

locate export platform FDI in L is not only a decision about how to serve L, but also a

 

7 Because of the economies of scale generated by the ﬁxed cost of investment, it will not be proﬁt
maximizing for the ﬁrm to invest in L for local sales only, and export to H from F.

50

decision about how to serve H. For the ﬁrm to prefer establishing export platform F DI in

L to exporting to both H and L from F requires:

(,_,L)[(1—02+(1—1>2_G]>{ll—ﬁrm],[tut-nix] (,5)

 

 

 

4 4 T 4 4
The left-hand side of (15) is the ﬁrm’s after-tax proﬁt from export platform F D1 in L. The
right-hand side of (15) is the ﬁrm’s proﬁt from exporting from F to H and facing a tariff
of rm and exporting from F to L and facing a tariff of ms.

Inequality (15) illustrates one of the interesting features of the model. Whereas
the preferences of H and L regarding the ﬁrm’s location decision are aligned in the
absence of a FTA (both H and L prefer for the ﬁrm to locate in L, provided the ﬁxed cost
of investment is not to large), H and L have opposing preferences about the ﬁrm’s
location decision in the presence of a FTA (L prefers to attract export platform F DI, but
H prefers for the ﬁrm not to locate in L). Not only do H and L have opposing preferences

about the ﬁrm’s location decision, but both H and L have policy instruments available to

inﬂuence the ﬁrm’s decision. Assuming G <G , L wants to attract export platform F DI,
so it will set rLF = 1 — f to discourage the ﬁrm from exporting to L from F and set tL as
high as possible while still satisfying (15) to extract the maximum possible tax revenue

from the ﬁrm’s investment. H, on the other hand, strictly prefers importing from F and
levying rHF e (7”,. , rHF*) to importing tariff-free from L, and thus wants to set rHF as low
as possible to violate (15) and encourage the ﬁrm to export to H from F. L will be

willing to lower tL as low as t, (G) to encourage the ﬁrm to locate export platform FDI in
L. H will be willing to lower rHF as low as 7m to encourage the ﬁrm to locate production

in F and export to H. Thus, in the presence of a FTA between H and L, when G < G,

51

there exists a “Bertrand-type” competition between L’s proﬁt tax rate, Q, and H’s tariff

rate on imports from F, rHF. Given this competition between H and L, (15) reduces to

(140F110+(I;I)-_G]Z[(1—f;7n) J (,6)

 

 

The left-hand side of (16) is identical to the left-hand side of (15), but the right-hand side

of ( 16) is equal to the right-hand side of (15) for er = FHF and rLF = 1 —f

Deﬁne Z(G) as the value of tL that satisﬁes (16) with equality for a given G. If tL
S t: (G), the ﬁrm will establish export platform FDI in L. If tL > t:(G), the ﬁrm will
export to both H and L from F. Thus, t—L (G) is the maximum tax rate for which the ﬁrm

at? (G)

will establish export platform FDI in L. Note that < 0. In other words, the larger

the ﬁxed cost of investment, the smaller the ﬁrm’s pre-tax proﬁt from export platform
FDI in L, and thus the smaller the tax rate necessary for the ﬁrm to be willing to establish
export platform FDI in L.

Given the competition between H and L to inﬂuence the ﬁrm’s location choice,
for L to prefer to attract export platform FDI requires the tax rate to be “big enough” for a

given G, as deﬁned by tL (G). For the ﬁrm to be willing to establish export platform FDI

in L requires L’s tax rate not to be “too big” for a given G, as deﬁned by t: (G).

Deﬁne Q such that:

gtgiilg) (17)
Q is the largest value of G for which L will be willing and able to attract export platform

FDI. The relationship between LL. (G), t—L(G), and Q is shown graphically in Figure 2.3.

52

IfG <Q , L will set (L = ;(G) and ru: = 1 -— f, and (15) will be satisﬁed for any
rm 6 (0, 1 —j), so the ﬁrm will establish export platform FDI in L. If, on the other hand,

~

G >g , H WIII SCI rHF = fHF E (rHF , rHF*), where

_ (1“!)2 (1‘1)2_ : (1_f_;m=)2
(1 t_,(G»[ 4 + 4 G] I 4 j (18)

 

 

The left-hand side of (1 8) is the ﬁrm’s after-tax proﬁt from export platform FDI in L,

facing a tax rate of tL (G). The right-hand side of (1 8) is the ﬁrm’s proﬁt from exports
from F to H facing rHF = PM, and exports from F to L facing rLF = 1 —f In other words, H

will set its tariff rate on imports to make the ﬁrm just indifferent between export platform

FDI in L and exports to both H and L from F. Thus, if G > Q , the ﬁrm will export to both
H and L from F.

Finally, deﬁne G such that:

_ (l—Ir (1—1)2_ =(1-f)’
(1 r_,(G)){ 4 + 4 G] I 9 j (19)

 

The left-hand side of (19) is the ﬁrm’s after-tax proﬁts from export platform F D1 in L,

facing a tax rate of t L (G). The right-hand side of (l 9) is the ﬁrm’s proﬁts from exports

from F to H facing r,,,: = r,,,:* and exports from F to L facing rLF = 1 — f. If G 2G , H will

be able to set er* and the ﬁrm will still choose to export from F to H. If G <G , H will

have to set fHF < rm:* to entice the ﬁrm to export from F to H rather than from L to H.

53

Proposition 2. In the presence of a FTA between H and L:

(a) if G 5g , H will set (er = rm, ), L will set (rip = 1 — f, tr = T, (G)), and the ﬁrm will
establish export platform FDI in L in equilibrium.

(b) if G e ( g , G ), H will set (rm: = r‘,,,. ), L will set (er = er’“, tL = r_,_ (G)), and the ﬁrm
will export from F to both H and L in equilibrium.

(c) if Ge [G ,G ), H will set (rHF = rim, L will set (rLF = rm, t, = r_,_ (G)), and the ﬁrm
will export from F to both H and L in equilibrium.

((1) if G 2 G , H will set (er = rm), L will set (er = rLF*, r, = 1), and the ﬁrm will
export from F to both H and L in equilibrium.

The ﬁrm’s equilibrium location decisions, both in the absence of a F TA between H and L

and in the presence of a FTA between H and L, are illustrated in Figure 2.4.

A surprising result, given Raff’s claim that a FTA cannot cause FDI destruction,

is that Q SQ ( 77,, ), which means that if Ge ( Q , Q ( 7m )), the ﬁrm will establish export

platform FDI in L in the absence of a FTA between H and L (since G <Q ( 7m )), but will
export to both H and L from F in the presence of a FTA (since G > Q ). Thus, a FTA may

cause FDI destruction.

Proposition 3. If Ge (Q , Q (7m )), a FTA between H and L will cause FDI destruction.

How does eliminating the tariff barrier between H and L make locating
production in L less attractive to the ﬁrm? It doesn’t; ceteris paribus, it makes locating

production in L more attractive to the ﬁrm. It also increases L’s willingness to attract

54

export platform FDI. So why, if a F TA makes locating production in L more attractive
than in the absence of a F TA, doesn’t a FTA cause FDI creation as in Raff (2004),
instead of F DI destruction? FDI destruction occurs because in response to economic
integration with L, H lowers its external tariff against imports from F, making locating
production in F instead of L more attractive to the ﬁrm than in the absence of a FTA. In

other words, H uses its trade policy to inﬂuence the ﬁlm’s location decision.

5. Conclusion

This chapter has shown that even if a country cannot itself attract FDI, it has
considerable power, through its trade policies, to inﬂuence a multinational ﬁrm’s location
decisions. Because a country’s optimal tariffs are positive in this model, economic
integration in the form of a FTA may generate competition between the low-cost
country’s tax rate and the hi gh-cost country’s external tariff rate to inﬂuence the ﬁrm’s
location decision. In some cases, the high-cost country may be able to “win” this
competition, in the sense that it is able to set its external tariff rate low enough to prevent
the ﬁrm from investing in the low-cost country. Because of this competition, a FTA may
cause FDI destruction, meaning that the monopolist operated a foreign afﬁliate in the
low-cost country in the absence of a FTA, but the competition between the integrating
countries in the presence of a FTA induces the ﬁrm to serve the FTA with exports.

This chapter leaves several interesting questions for future investigation. For
example, what are the welfare effects of economic integration in the form of a FTA for

the integrating countries? Does this model provide new insights into the trade creation

55

and trade diversion effects of economic integration? What about the relationship between
bilateral trade liberalization and multilateral trade liberalization? between trade
liberalization and investment liberalization? These are all potential areas for future

research using this relatively simple model.

56

Figure 2.1

A Graphical Illustration of 7H,

1: 0.2 f= 0.25

 

0.15

0. I
fromF( r HI”)

fromL( r)
fromF( r)

 

0.05

 

 

 

 

 

fromL(r) is H’s social welfare from importing from L and levying rHL

fromF(r) is H’s social welfare from importing from F and levying mp

57

Figure 2.2

A Graphical Illustration of 7m

 

     

 

 

 

 

 

1= 0.2 f= 0.25
I I I I I
0.15
0.1 - 'I
O . .I . I
.. l - .
fromL( r) - 17’:- - - - - - -: ------ z": """""""" fromL(O)
. I '
fromF(r) : :
..... | I
I l
l l
0.05 ' ' 7
I I
I l
I 1
I l
l I
l I
l l
l I
l l
I I
0 I I
l l 4 I
~ r .1
rHF 0.2 HI 0.4 . 0.8
r 0.9

 

fromL(r) is H’s social welfare from importing from L and levying rHL

fromF(r) is H’s social welfare from importing from F and levying mp

58

 

Figure 2.3

A Graphical Illustration of Q

1: 0.2 f= 0.25

 

 

 

  

   

 

 

 

0.7
0.6
tLupper(G)0.4 3 _
t Llovver(G ) 3
0.2 _
I ..o
0.. |
c.000.0000000000'. '
O I
l 1 I r l
0
O 0.05 0'] (“5 Q 0.2 0.25
0 G 0.3

tLuppc,(G) is 17 (G), the highest tax rate for which the ﬁrm will invest in L

tLiowcr(G) is t ,1 (G), the lowest tax rate for which L prefers to attract investment

59

Figure 2.4

Firm’s equilibrium location decisions

 

No F T A: locate production in L I locate production in F
(export platform FDI) : (export to both H and L)
increasing G —-* Q (7111, )
G: I A ‘ I
F T A: locate production in L locate production in F

(export platform FDI) (export to both H and L)

-----IIQ -

 

_. FDI destruction

Proposition 3. If Ge (Q , Q ( 77m )), a FTA between H and L will cause FDI destruction.

60

CHAPTER THREE: PREFERENTIAL TRADE AGREEMENTS AND TAX

COMPETITION WITH INTERNATIONALLY MOBILE FIRMS

1. Introduction

The connections between economic integration and tax competition have received
considerable attention in recent years. Much of this attention is attributable to policy
debates surrounding the formation and expansion of the European Union. Speciﬁcally, it
is argued that economic integration has increased international capital mobility within the
EU and may generate a “race to the bottom” in corporate income tax rates, as EU
members compete for mobile capital. In response to such concerns, serious debate has
emerged on the need for tax harmonization in the EU (see for example Sim 1990,
Cnossen 2003, McLure 2005).

In the standard models of tax competition for mobile capital (see Wilson 1999), a
lower marginal tax rate levied by one region attracts capital to that region from other
regions, thereby generating a negative ﬁscal extemality (smaller tax base) for other
regions. As a result, equilibrium tax rates are inefﬁciently low (and if taxes are used to
ﬁnance local public goods, local public goods are underprovided). This result illustrates
one concern surrounding the effect of economic integration on tax rates in the EU.

Devereux et al. (2002) present several stylized facts about the evolution of taxes
in EU and G7 countries and claim that standard models of tax competition are inadequate
to explain these stylized facts. They write, “The view that corporate income tax rates

have fallen in response to increased mobility of capital, as countries compete to lower the

61

cost of capital within their own jurisdictions, is therefore not generally borne out by the
data. An alternative possibility is that countries may instead compete for the activities of
mobile multinational ﬁrms,” (Devereux, et al. 2002, p. 542-543).

Zodrow (2003) reviews the tax competition literature in light of the debate over
tax harmonization in the EU. He notes, “the basic tax competition model assumes a ﬁxed
capital stock in the union. This assumption is of course unwarranted to the extent that the
union faces an elastic supply of capital,” (Zodrow 2003, p. 656). In other words, standard
tax competition models fail to incorporate competition for foreign direct investment
(FDI) from outside the competing regions.

This paper develops a model of tax competition that incorporates competition for
FDI from outside the competing regions. Two identical countries each have a ﬁxed
supply of labor, which produces a freely tradable intermediate good. This intermediate
good is transformed into a ﬁnal good by a ﬁxed number of domestic ﬁrms, each of which
earns positive proﬁts. While the number of domestic ﬁrms in each country is ﬁxed, the
two identical countries face an elastic supply of ﬁrms from the rest of the world. If a ﬁrm
from the rest of the world locates in one of the two countries, it is subject to that
country’s taxes, and repatriates its after-tax proﬁts in units of the intermediate good.
When there is no trade in the ﬁnal good between the two identical countries or the rest of
the world (“autarky” in this model), both countries choose a positive equilibrium taxes to
exercise market power in the “market” for internationally mobile ﬁrms by taxing away
some of the proﬁts earned by ﬁrms locating within their borders. However, when the two
identical countries form a preferential trade agreement (PTA) that allows trade in the

ﬁnal good between the two countries, but not with the rest of the world (that is, the two

62

identical countries move from “autarky” to “free trade”), tax competition for
intemationally-mobile ﬁrms from the rest of the world ensues, driving equilibrium taxes
in both countries to zero.

The model presented in this chapter attempts to clarify the relationships between
economic integration, tax competition, and foreign direct investment by multinational
ﬁrms. Section 2 describes the model in more detail. Section 3 derives each country’s
equilibrium taxes in autarky. Section 4 considers integration in the form of a preferential

trade agreement and its effect on equilibrium taxes. Section 5 concludes.

2. The Model

The basic setup of the model is similar to Krugman (1980). Firms in two identical
countries, H and F, each produce a single variety of a differentiated ﬁnal good.
Equilibrium is analyzed both in the absence of trade in this ﬁnal good (autarky), and
when these two countries freely trade this ﬁnal good with each other. This paper extends
Krugman’s model in three signiﬁcant ways. First, whereas Krugman (1980) assumes free
entry of domestic ﬁrms in each country, which serves to drive equilibrium proﬁts of each
ﬁrm to zero, this paper assumes that the residents of H and F each own a ﬁxed number of
domestic ﬁrms that is smaller than the number of ﬁrms that would exist if there were free
entry, and thus each ﬁrm earns positive economic proﬁts. Second, this paper introduces a
third country, R (the rest of the world). Residents in R also own a ﬁxed number of ﬁrms.
Since the number of ﬁrms in each country is ﬁxed and each ﬁrm earns positive economic

proﬁts, this opens up the possibility for foreign direct investment (FDI) in each country.

63

If internationally mobile ﬁrms owned by residents in R can earn a higher proﬁt by
locating in either H or F than they could from locating in R, they will do 50. Third,
whereas in Krugman (1980), labor is used to produce the differentiated product directly,
in the current model, labor produces a freely tradable intermediate good using constant
retums-to-scale technology. This intermediate good, then, is transformed by ﬁrms into
differentiated ﬁnal goods. The intermediate good serves an important role in this model.
While ﬁrms are internationally mobile in this model, ﬁrm owners and labor are not. If
ﬁrms owned by residents in R locate in either H or F and earn proﬁts there, these ﬁrms
must be able to repatriate their proﬁts to the ﬁrm owners in R (who do not move with the
ﬁrms to H or F). In the model, the intermediate good serve as the numeraire, and proﬁts
of R-owned ﬁrms are repatriated in units of the freely-tradable intermediate good.

Since H and F are identical, focus on the behavior of a representative consumer in
H. A representative consumer derives utility from consuming varieties of the

differentiated ﬁnal good as represented by the utility function

n 19
U(c,,..., c,,)= 21:6: (1)

where c, is the representative consumer’s consumption of variety i, n is the number of
varieties and is assumed to be large, and 6 < 1.The representative consumer maximizes

her utility subject to a budget constraint

: pici S I (2)
i=1

where p,- is the price of variety i and I is the representative consumer’s income.

64

Solving the representative consumer’s constrained optimization problem yields
the demand for each variety i:

pi G:
ci=1 ? (3)

where

n 6 6—1
F = [Z 1),-0"] (4)
i=1

is an aggregate price index. A detailed derivation of (3) is provided in Appendix 1.

Given the demand for each variety (3), we can analyze the proﬁt-maximizing
behavior of ﬁrms in H to determine the price and quantity of each good. Labor (L) is used
to produce a freely tradable intermediate good y according to the constant-returns-to-
scale technology:

L = y (5)
Normalize the total size of the population/labor force (L) in each country to 1 and let the
price of the intermediate good be the numeraire. This, in turn, normalizes the wage rate
to 1.

Each ﬁnal good producing ﬁrm faces a ﬁxed cost of production of a units of the
intermediate good and a marginal cost of production of ,8 units of the intermediate good.
In other words, the cost of producing x,- units of the ﬁnal good is:

yr 2 a + ﬁx,- (6)
With it identical ﬁrms, the market for the intermediate good clears when each ﬁrm
receives l/n units of the intermediate good. This ﬁxes the output of each ﬁrm for a given

number of ﬁrms.

65

__1—an

xi —
,Bn

 

(7)

Solving the following proﬁt maximization problem faced by each ﬁrm will

determine the equilibrium proﬁt-maximizing price set by each ﬁrm.

man prxr — yr

Having normalized the size of the population to 1, the ﬁnal goods market clears when the
quantity of each variety demanded by the representative consumer equals the quantity of
output by each ﬁrm, or c,- = x,. Solving the ﬁrm’s proﬁt maximization problem yields the

proﬁt-maxmizing price for each variety, which is a constant markup over marginal cost:

.= =2
pi p 6 (8)

A detailed derivation of (8) is given in Appendix 2.
Since the equilibrium price for each ﬁrm is the same (p,- = p), the representative
consumer’s demand for each variety (3) reduces to:

c,-= 1- (9)

”P
That is, the representative consumer will spend a fraction l/n of her income on each
variety. Substituting the equilibrium price (8) into the representative consumer’s demand

(9) yields the proﬁt maximizing quantity produced by each ﬁrm (since c,- = x,):

61

xi: 3,-1- (10)

Since the price of each variety of the differentiated ﬁnal good (8) is the same, and
each variety enter’s the representative consumer’s utility function (1) symmetrically, the

representative consumer’s indirect utility function is given by:

66

V=n[%) (11)

That is, the representative consumer consumes an equal quantity (8) of each of n varieties
of the differentiated ﬁnal good.

To this point, we have neither speciﬁed nor solved for a value of n. To begin, let’s
derive the indirect utility-maximizing value of n. In order to do this, we need to derive the
representative consumer’s income as a value of n. The representative consumer’s income
consists of wage income (recall that the wage rate equals 1 due to the constant-retums-to-
scale technology used to produce the freely tradable intermediate good, which serves as
the numeraire) plus any proﬁts generated by domestic ﬁrms. In other words, the
representative consumer’s income is given by:

I = 1 + 7m (12)
where it is the proﬁt earned by each of n domestic ﬁrms. Since we know the price (8),
quantity produced (9), and cost of production (6) for each ﬁrm, we can derive it as a
function of n.

It: (l—t9—orn)
9h

(13)
First, note that the proﬁt of each ﬁrm is decreasing in the number of ﬁrms. This is
due to internal economies of scale. Because of the ﬁxed costs of production and the
limited supply of the intermediate good, the more ﬁrms there are, the smaller each ﬁlm
is, moving each ﬁrm “up” its average total cost curve and reducing its proﬁts. Also note

that the zero-proﬁt (free-entry) level of n is

n”: E (14)
a

67

Substituting the proﬁt of each ﬁrm (13) into the representative consumer’s
income (12) yields the representative consumer’s income as a function of the number of
ﬁrms:

: (l-Cm)
6

 

I (15)

Substituting this back into the representative consumer’s indirect utility function (11)

gives the representative consumer’s indirect utility as a function of the number of ﬁrms:

in

Maximizing (16) with respect to it yields the indirect utility-maximizing number of ﬁrms:

 

V

n=3_—9=n0 (17)
a

In other words, free entry will guarantee the efﬁcient number of ﬁrms. This is a standard
result in the Krugman (1980) model, which assumes an iso-elastic utility function
identical to (1) and free entry of domestic ﬁrms.

In this model, however, we assume that the number of domestic ﬁrms in H is

ﬁxed below the number that would exist with free entry. That is, consider n <-1———6. In
a

this case, domestic ﬁrms in H are earning positive economic proﬁts, since from ( 12), if

n <l_—6, then it > 0. If the proﬁt earned by domestic ﬁrms in H is greater than the proﬁt
or

available in R (the rest of the world), internationally mobile ﬁrms from R have an
incentive to locate in H. Attracting ﬁrms from R would have several important
consequences for H. In autarky there is no trade in the ﬁnal good, so the only way for

consumers to access additional varieties is to attract internationally mobile ﬁrms from R.

68

Thus, the entry of ﬁrms from R would increase the number of domestically available
varieties, which beneﬁts consumers who are currently consuming an inefﬁciently low
number of varieties. On the other hand, the entry of ﬁrms from R would reduce the
proﬁts of domestic ﬁrms (and in turn, the incomes of domestic consumers). Either way,
the prospect of attracting ﬁrms from R opens the door for ﬁscal policy on the part of the
government of H. The government of H may be willing to subsidize ﬁrms to increase the
number of domestically available varieties, or may wish to tax ﬁrms to exercise market

power over the proﬁts that ﬁrms from R require to locate in H.

3. Equilibrium in Autarky

Consider H and F in autarky. That is, although the intermediate good is freely

tradable between H, F and R, these countries do not engage in trade in the ﬁnal good. H

and F each have n < 1:0— domestic ﬁrms. Again, since H and F are identical, we will
a

focus on H. Each ﬁrm in H is earning proﬁts according to equation (13). In the rest of the
world (R), ﬁrms earn a proﬁt given by 7rR(N+N*), where N is the number of R-owned
ﬁrms that locate in H and N * is the number of R-owned ﬁrms that locate in F. Although
this generic proﬁt function is exogenous to this model, a speciﬁc proﬁt function with the
properties of this function could be generated by a more speciﬁc model of the rest of the
world. Just like the proﬁts of domestic ﬁrms decrease as the number of ﬁrms increases,
the proﬁts of ﬁrms in R decreases with the number of ﬁrms in R. Thus, the more R-
owned ﬁrms that locate in H or F (the fewer ﬁrms that remain in R), the higher the proﬁts

of the remaining R ﬁrms will be. In other words, JTR '> 0.

69

In order to inﬂuence the number of R-owned ﬁrms that locate in H, the
government of H may levy a tax (subsidy) of t > 0 (t < 0) on ﬁrms that locate within its
borders. To see the effects of a such a tax, ﬁrst note that in equilibrium, the proﬁt level in
H (net of the tax) must equal the proﬁt level in the rest of the world, 7rR(N+N*). (Treat the
policy of the government of R as exogenous.) This generates the following relationship
between the size of the tax and the size (output) of each ﬁrm as follows:

Px- ((1 +3304 = 7rlt(N+N’") (18)
Equation (18) is an equal proﬁt condition. It says that the after-tax proﬁt of ﬁrms locating
in H must equal the proﬁt of ﬁrms remaining in R. Substituting'the proﬁt-maximizing
price chosen by each ﬁrm (8) into (18) and solving for x and determines a relationship
between the size of (quantity produced by) each ﬁrm (x), and the tax chosen by the

government of H (t).

x = 6(7rR(N+N*)+a+t)
(1-9)ﬂ

E: 6 [”R'flﬂ+1] (20)
dt (1—19)/3 dt -

(19)

 

 

Next, we need to determine the relationship between the tax and the number of R-
owned ﬁrms choosing to locate in H. A cost to H of attracting R-owned ﬁrms is that R-
owned ﬁrms use up domestic resources in H. Speciﬁcally, proﬁts earned by R-owned
ﬁrms are repatriated to ﬁrm owners in R in units of the intermediate good. This generates
the following intermediate good market clearing condition:

(a +,Bx)(N+n)=1—N-7rR (21)
The left-hand side of (21) shows the total intermediate good used by ﬁrms locating in the

country (11 domestic ﬁrms plus N R-owned ﬁrms), and the right—hand side shows the total

70

intermediate good available, after paying in; for N foreign ﬁrms. Solving equation (21)
for x determines a relationship between the size (output) of each ﬁrm (x), and the number

of R-owned ﬁrms that locate in H (N).

x= 1—a(N+n)—N7rR

,8(N + n) (22)

 

Equation (18), or equivalently equation (19), is an “equal proﬁt” (not zero proﬁt)
condition, which determines a relationship between x and t. Equation (21), or
equivalently equation (22), is an intermediate good market clearing condition, which
determines a relationship between x and N. Combining equation (22) with equation (19)
determines the relationship between the tax chosen by the government of H (t) and the

number of R-owned ﬁrms that locate in H (N).

(l —6)(l +n7rR) =
(N+n)

 

7r, + a + 61 (23)

EN=_ t§I(N+n)2 (24)
dt (1—6)(1+n7rR)+(N+n)(N+t9n)7rR'

 

Equation (24) shows that as the tax levied by H increases, the number of R-owned ﬁrms

that will locate in H falls, or (id—1:,< 0.

To determine the equilibrium level of the tax in H in autarky, the government of
H maximizes the representative consumer’s indirect utility function with respect to the
tax. Solving this problem yields the following optimal tax formula for H in autarky:

t = N- 7rR' (25)
The details of the derivation of (25) are provided in Appendix 3. The result in equation

(25) is the familiar terms-of-trade effect. If 7rR(0) < it, then N > 0 (ﬁrms will relocate from

71

R to H). Since H must pay proﬁts to R—owned ﬁrms that locate in H, H is effectively
“importing” ﬁrms from R at a “price” of it per ﬁrm. Since H can inﬂuence the price of
these imports (R-owned ﬁrms’ after-tax proﬁts) through its choice of tax, H sets a
positive “optimal tariff” (equilibrium tax). In other words, H uses a positive tax to
exercise market power over the proﬁts that R-owned ﬁrms must receive to locate in H.
Because H and F are identical, they face identical problems with respect to the
entry of R-owned ﬁrms in autarky. Since H and F are identical, F will choose an
equilibrium tax given by the formula t* = N *-7rR’ where N * = N and therefore t* = t.

Thus, both H and F will choose positive taxes in autarky.

4. Equilibrium in a Preferential Trade Arrangement

What happens, however, if H and F form a preferential trade arrangement, but
remain closed to ﬁnal goods trade with R? That is, both the intermediate good and the
differentiated ﬁnal good are freely tradable between H and F, but only the intermediate
good is tradable with R.

The following two conditions must be true in an equilibrium with ﬁrms in each
country:

i) px—(a+ﬂx)—t=7rR(N+N*)

ii) p*x*—(a +ﬂx*)—t* =7rR(N+N*)

Condition i) says that given that N R-owned ﬁrms locate in H and N * R-owned ﬁrms
locate in F, the proﬁts in H (net of the tax) must equal proﬁts in the R. Condition ii) is

equivalent for F.

72

Two facts are important to note. First, because there is free trade in the ﬁnal good,
the demand curve faced by each ﬁrm in the PTA is identical, regardless of the ﬁrm’s
choice to locate in H or F. Second, also because of free trade, the price of the
intermediate good (the numeraire) is equalized in the PTA. Given these facts, each ﬁrm
locating in the PTA, regardless of the country in which it locates , will choose the same
price (p = p*) and the same level of output (x = x*) to maximize proﬁts. Thus, we can
rewrite i) and ii) as:

i’) px-(a +,8x)—t=7rR(N+N*)

ii’) px — (a + ,Bx) -— t* = 7rR(N + N*)

Clearly if both i’) and ii’) are satisﬁed, t = t* in an equilibrium with preferential free
trade and ﬁrms operating in both countries. Consider the alternatives. If t > t*, the proﬁts
for ﬁrms locating in F are greater than the proﬁts for ﬁrms locating in H, and thus any
ﬁrm currently operating in H will relocate production to F. With no ﬁrms remaining in H,
condition i’) is irrelevant. Similarly, if t < t*, all ﬁrms will locate in H, and ii’) is
irrelevant.

Consider for a moment that t = t* = N- n' as in autarky. Is this a Nash equilibrium
in taxes, or could the government of either H or F increase the indirect utility of the
representative consumer in its country by deviating?

With preferential free trade, the indirect utility of the representative consumer in

H is:

 

191 0
V: at: :1:
(N+n+N +n)[,B(N+n+N*+n*)I (25)

That is, the representative consumer in H is able to enjoy not only the N + n domestically

produced varieties, but also the N * + n * varieties produced in F. What happens to the

73

number of varieties available to consumers in H if the government of H lowers its tax
slightly below t = N 7r' ? Nothing! While a reduction in taxes may inﬂuence the
distribution of ﬁrms within the PTA, it does not affect the number of varieties available
to consumersg. Note that regardless of the location of ﬁrms within the PTA, consumers in
both H and F will be able to consume the varieties of ﬁnal good produced by all ﬁrms in
the PTA. That is, all F D1 in the PTA will be export platform FDI, whereby the ﬁrm not
only serves consumers in the host country, but also exports its good to consumers in the
other country in the PTA.

To say that a reduction in a country’s tax does not affect the number of varieties
available to consumers is not to say that the reduction in taxes has no effect on
consumers. The representative consumer’s indirect utility is a function not only of the
number of available varieties, but also of the representative consumer’s income, which
depends on the number of ﬁrms locating in that country (the country’s tax base):

I =1 + n-er(N+ N*) + t-(N + n) (26)
While a small reduction in t given t* does not have a signiﬁcant effect on the number of
varieties available to consumers in H, it does have a signiﬁcant effect on the number of
ﬁrms that actually locate in H. As described above, if H sets I < t*, then the N * + n*
ﬁrms that previously operated in F will relocate to H, transforming equation (26) into

[=1+n'7rR(N+N*)+t'(N+n+N*+n*) (26’)
That is, a small reduction in t results in a doubling of the tax base in H, and thus an

increase in the income that the representative consumer in H can spend on the available

 

8 Actually, at this lower tax, some additional ﬁrms from R may now locate in H, but assume that for a small
reduction in t, this is negligible.

74

varieties. This provides a strong incentive for the government of H to set t slightly below
t*.

Since H and F are identical, the government of F’s incentive to set t* slightly
below t is equally strong. H and F will therefore compete in taxes a Bertrand-type
fashion. This competition will cease when t = t* = 0. If t* = 0, it is straightforward to
show that H’s best response is to set t = 0, since for t > 0, all ﬁrms in the PTA will locate
in F and H will not collect any positive tax revenue, but for t < 0, all ﬁrms in the PTA
will locate in H, but the representative consumer’s income is lower than if t were equal to
zero. Thus, when H and F form a PTA, the Nash equilibrium in taxes is t = t* = 0.

This result is reminiscent of J aneba (1 998), wherein it is shown that in a Brander
and Spencer (1985)-type model of two national oligopolists competing in a third market,
each government has an incentive to subsidize its national ﬁrm in the absence of ﬁrm
mobility, resulting in wasteful subsidy competition. However, when ﬁrms are
internationally mobile, tax competition eliminates these wasteful subsidies. J aneba (1998)
starts with free trade and shows that introducing international ﬁrm mobility eliminates
government subsidies that would be chosen in the absence of ﬁrm mobility. In other
words, ﬁrm mobility in the presence of free trade induces tax competition between
competing regions. In contrast, the present model starts with international ﬁrm mobility
and no trade in the ﬁnal good, and shows that introducing preferential free trade
eliminates government taxes that would be chosen in the absence of trade. In other words,
preferential free trade in the presence of ﬁrm mobility induces tax competition between

competing regions.

75

Despite these similarities, the welfare effects of tax competition in the two models
stand in stark contrast. In Janeba (1998), the subsidies chosen in the absence of
international ﬁrm mobility are welfare-reducing relative to the laissez-faire equilibrium
generated by introducing ﬁrm mobility. That is, tax competition is welfare-improving. In
the current model, the taxes chosen by the governments of H and F in the absence of a
preferential trade agreement between the two countries are optimal (indirect utility
maximizing) for each country. The tax competition induced by introducing trade in the
ﬁnal good between H and F generates a prisoner’s dilemma for H and F; although
retaining t = t* > O as in autarky generates a higher level of indirect utility for the
representative consumers in H and F than the Nash equilibrium t = t* = 0, each
government has an incentive to set a tax slightly lower than the other country to expand
its tax base at the expense of the other country. Thus, in the presence of a FTA between H
and F, when H and F set t = t* = 0, more R-owned ﬁrms locate in the PTA than would

have in the absence of a F TA, which is welfare-worsening for both H and F.

5. Conclusion

The model presented in this chapter extends existing models of economic
integration and tax competition by incorporating internationally mobile ﬁrms from
outside the integrating region. Whereas in standard tax competition models the motive for
setting positive taxes is to ﬁnance a public good, the current model introduces an
alternative motive for a country to set a positive tax: to exercise market power over the

returns that internationally mobile ﬁrms require to locate in that country. Thus, in

76

autarky, governments set positive tax rates in equilibrium to lower the after-tax proﬁts of
ﬁrms locating within their borders, thereby limiting the entry of internationally mobile
ﬁrms from the rest of the world. When two countries allow preferential free trade in the
ﬁnal good, both governments have incentives to lower their taxes to increase their tax
bases, since they do not internalize the ﬁscal extemality generated by their action. Thus,
tax competition eliminates the positive taxes chosen by each country in autarky.

The intuition here is relatively straightforward. When two countries face an
upward-sloping supply of foreign ﬁrms, the optimal tax is positive, since each country
can use positive taxes to inﬂuence its “terms-of-trade” in ﬁrms (increasing the tax on
ﬁrms effectively lowers the “price” a country must pay for “imported” ﬁrms)9. When the
two countries integrate, however, each country can discontinuously increase the number
of ﬁrms it attracts by lowering its tax by a small amount. In other words, preferential free
trade introduces tax competition for internationally mobile ﬁrms from the rest of the

world, which eliminates the “terms-of-trade” motivations for positive taxes in autarky.

 

9 Note that this result depends crucially upon the assumption of market power of the two countries. From
(25), if countries do not have market power in the “market” for intemationally-mobile fums (if 71" = 0),
then countries do not have an incentive to set positive taxes in autarky.

77

APPENDICES

78

1. Deriving the representative consumer’s demand for each variety

The ﬁrst-order condition of the consumer’s constrained optimization problem is:
66,6" — zip,- = 0

where )t is the Lagrangian multiplier on the consumer’s budget constraint.
0,-0" = lpﬂl

(picaﬂ-I ____ Apr-69]

n 6, 9-1
DeﬁnezP = [21710—1]

6-1
A: 6_
p

 

Substituting this back into the ﬁrst-order condition yields:

916—]
0—1 I
C" _ — p16—

1
pi I?
Cir-I —I——)—

 

79

2. Deriving a representative firm’s profit-maximizing price

A ﬁrm’s proﬁt maximization problem is given by:

max

x,

pixi ‘3’1‘
Substituting for p,- using the ﬁrst-order condition from the representative consumer’s

constrained optimization problem (see Appendix 1) and remembering that x,- = c,- in

equilibrium (since L = 1), the ﬁrm’s proﬁt maximization problem becomes:

ma (gjx? _ (a + ﬂxi)

x,-

The ﬁrst-order condition of the ﬁrm’s proﬁt maximization problem is:

80

3. Deriving the equilibrium tax in autarky
The representative consumer’s indirect utility function in autarky is:
V = (N+n)x9

Maximizing this with respect to the tax levied by H (t) yields:

ﬂ: c—ﬂixg + 0(N+n)x6'lg = 0
dt dt dt
£1’ﬂx = -19(N+n) £11

dt dt

dN __ 62(N+n)( ,dN )
—x—-——— e,—+1
dt (1—6),B dt

 

(1N [H 62(N+n)7rk'] :_62(N+n)

T (1-6)ﬂ (1-6’)/3

 

 

dN [york +a+t) +192(N+n)7rk'] :_62(N+n)

T (1-9)ﬂ (1-9)ﬂ (1-6)ﬂ

-C:l—N— (HR +a +t +I9(N+ n)7rR')= -6(N+n)
t

d_N: 6(N+n)
dt nR+a+t+t9(N+n)7rR'

 

6(N + n)2 = 6(N + n)

 

 

 

- - [from(24)]
(1—6)(1+n7rR)+(N+n)(N+6n)rrR' 72,, +a+t+l9(N+n)7rR'
7r, +a+t+e(N+n)rr,,'= (1—6)(1+nﬂk)+(N+9n)7TR'
(N+n)
7rR+or+t+l9(N+n)7rR’=nR+a+9t+(N+t9n)7rR' [from(23)]

(1-9)t = (1-6)N7r,,'

t=N-7rR'

81

REFERENCES

82

REFERENCES

Baltagi, Badi H., Peter Egger, and Michael Pfaffermayr, 2005. Estimating Models of
Complex F DI: Are There Third-Country Effects? Working Paper. Texas A&M
University, University of Munich, and University of Innsbruck.

Bernard, Andrew B., J. Bradford Jensen, and Peter K. Schott, 2005. Importers, Exporters,
and Multinationals: A Portrait of Firms in the US that Trade Goods. Working
Paper 05-10. Peter G. Peterson Institute for International Economics.

Blonigen, Bruce A., 2005. A Review of the Empirical Literature on FDI Determinants.
Working Paper #11299. National Bureau of Economic Research.

Blonigen, Bruce A., Ronald B. Davies, Glen R. Waddell, and Helen T. Naughton, 2004.
FDI in Space: Spatial Autoregressive Relationships in Foreign Direct Investment.
Working Paper #10939. National Bureau of Economic Research.

Brainard, S. Lael, 1997. An Empirical Assessment of the Proximity-Concentration
Trade-off Between Multinational Sales and Trade. American Economic Review
87, 520 — 544.

Brander, J .A. and B.J. Spencer, 1985. Export Subsidies and International Market Share
Rivalry. Journal of International Economics 18, 83 — 100.

Brander, James A. and Barbara J. Spencer, 1987. Foreign direct investment with
unemployment and endogenous taxes and tariffs. Journal of International
Economics 22, 257 -— 279.

Carr, David L., James R. Markusen, and Keith E. Maskus, 2001. Estimating the
Knowledge-Capital Model of the Multinational Enterprise. American Economic
Review 91, 693 — 708.

Caves, Richard E., 1982. Multinational enterprise and economic analysis. Cambridge
UP, Cambridge.

Cnossen, Sijbren, 2003. How much Tax Coordination in the EU? International Tax and
Public Finance 10, 625 — 649.

Cook, Nathaniel PS, 2006. Comment on “Preferential trade agreements and tax

competition for foreign direct investment”. Manuscript. Michigan State
University.

83

Davis, Donald R. and David E. Weinstein, 2001. Market Size, Linkages, and
Productivity: A Study of Japanese Regions. Working Paper #8518. National
Bureau of Economic Research.

Devereux, Michael P., Rachel Grifﬁth, and Alexander Klemm, 2002. Corporate income
tax reforms and international tax competition. Economic Policy 17, 450 — 494.

Ekholm, Karolina, Rikard Forslid, and James R. Markusen, 2003. Export Platform
Foreign Direct Investment. Working Paper #9517. National Bureau of Economic
Research.

Gronau, Reuben, 1974. Wage Comparisons — A Selectivity Bias. Journal of Political
Economy 82,1119 — 1143.

Hanson, Gordon H., Raymond J. Mataloni, Jr., and Matthew Slaughter, 2001. Expansion
Strategies of US Multinational Firms, in Collins, SM. and Dani Rodrik (Eds),
Brookings Trade Forum: 2001. Brookings Institution Press, Washington DC,
245 — 294.

Hanson, Gordon H., Raymond J. Mataloni, Jr., and Matthew Slaughter, 2003. Vertical
Production Networks in Multinational Firms. Working Paper #9723. National
Bureau of Economic Research.

Harris, Chauncey D., 1954. The Market as a Factor in the Localization of Industry in the
United States. Annals of the Association of American Geographers 44, 315 — 348.

Head, Keith and Thierry Mayer, 2004. Market Potential and the Location of Japanese
Investment in Europe. Review of Economics and Statistics 86, 959 — 972.

Heckman, James J ., 1979. Sample Selection Bias as a Speciﬁcation Error. Econometrica
47, 153 — 161.

Helpman, Elhanan, 1984. A Simple Theory of International Trade with Multinational
Corporations. Journal of Political Economy 92, 451 — 471.

Helpman, Elhanan, and Paul R. Krugman, 1985. Market Structure and Foreign Trade:
Increasing Returns, Imperfect Competition, and the International Economy. MIT

Press, Cambridge, MA.

Helpman, Elhanan, Marc Melitz, and Yona Rubinstein, 2004. Trading Partners and
Trading Volumes. Manuscript. Harvard University.

Hwang, Hong and Chao-cheng Mai, 2002. The tariff-j umping argument and location
theory. Review of International Economics 10, 361 — 368.

84

Motta, Massimo and George Norman, 1996. Does Economic Integration Cause Foreign
Direct Investment? International Economic Review 37, 757 — 783.

J aneba, Eckhard, 1998. Tax competition in imperfectly competitive markets. Journal of
International Economics 44, 135 — 153.

Krugman, Paul, 1980. Scale Economies, Product Differentiation, and the Pattern of
Trade. American Economic Review 70, 950 - 959.

Markusen, James R., 1984. Multinationals, Multi-Plant Economics, and the Gains from
Trade. Journal of International Economics 16, 205 — 226.

Markusen, James R. and Anthony J. Venables, 2000. The Theory of Endowment, Intra-
Industry, and Multi-national Trade. Journal of International Economics 52, 209 —
234.

Motta, Massimo, 1992. Multinational ﬁrms and the tariff-jumping argument. European
Economic Review 36, 1557 — 1571.

Motta, Massimo and George Norman, 1996. Does economic integration cause foreign
direct investment? International Economic Review 37, 757-783.

McLure, Charles E., J r., 2005. The European Commission’s Proposals for Corporate Tax
Harmonization. CESifo Forum, 32 — 41.

Neary, J. Peter, 2002. Foreign Direct Investment and the Single Market. Manchester
School 70, 291 — 314.

Raff, Horst, 2004. Preferential trade agreements and tax competition for foreign direct
investment. Journal of Public Economics 88, 2745 — 2763.

Shatz, Howard J. 2004, US Multinational Afﬁliate Exports from Developing Countries.
Journal of Economic Geography 4, 323 — 344.

Sinn, Hans-Wemer, 1990. Tax Competition and Tax Coordination in Europe. European
Economic Review 34, 489 — 504.

United Nations Conference on Trade and Development (UNCTAD), 2000. Tax
Incentives and Foreign Direct Investment: A Global Survey. ASIT Advisory
Studies 16. United Nations, New York, NY.

Vousden, Neil, 1990. The economics of trade protection. Cambridge: Cambridge
University Press, 122-126.

Wilson, John D., 1999. Theories of Tax Competition. National Tax Journal 52,
269 — 304.

85

Wooldridge, Jeffrey M., 2002. Econometric Analysis of Cross Section and Panel Data.
MIT Press, Cambridge, MA.

Yeaple, Stephen R., 2003. The Role of Skill Endowments in the Structure of US Outward
Foreign Direct Investment. Review of Economics and Statistics 85, 726 — 734.

Zodrow, George R., 2003. Tax Competition and Tax Coordination in the European
Union. International Tax and Public Finance 10, 651-671.

86

   

ililtiliiiiiiiiilii(iii