r4 Laﬁ-

1 ""335. .. c

”33.... ,
mg“ ﬂaw
%§un’i2
:3

a .. .
. .r. Tn‘ ,
.333 wan. f

y ‘.
z: u» r. .u ..
a :H gwmn
z. . . ....,
3 4% .3
gums .
25$

,F'L'Effﬁm

:1... ..
2%.: a.
5

 

 

 

“In...
36053.11?
12%... f. . 51.1..

 

5.x.
1. I .
.mﬂmhavﬁ. ..
, . inﬂux.)

’ a
mu 1:.
.5} ﬁaﬁﬂ.
I.»
$4.. (any.
.225... 4....
3.... a

..

. ,.
a
n.3,",

t a
5..

Utgi
(3.2.4... uni-{H41 .14x

"1...". 4.. pug“

o #3. in? .Hﬁssg
‘ I’LIICHJﬁhhuzxﬁz
21.91....”

.1 c ,. ., .
131.31; align); .. .
v, «1113...... ...
ii 3.2.;
v)

gunk: a.”
ﬂu.) 1‘3.)

Hun.

 

l a...“
_ at... 3%:
:12“ ﬁg“... , _.........n,:aa..... . _
, 3.". ,

..
.~.v

 

 

rﬂfls

This is to certify that the
dissertation entitled

A THEORETICAL AND EMPIRICAL INVESTIGATION OF
THE MULTINATIONAL FIRM’S CHOICE OF ENTRY MODE

presented by
Bong Geul Chun

has been accepted towards fulﬁllment
of the requirements for the

Ph. D. degree in Economics

 

 

ffww/W'

Major Pfofessor’s 8 nature
M

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

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

 

JUN. 202

007

 

i " f. ‘
Ii . I.
V .l' U .

GT

 

 

 

 

 

 

 

 

 

 

 

 

2/05 plelRC/DateDue.indd-p.1

 

A THEORETICAL AND EMPIRICAL INVESTIGATION
OF THE MULTINATIONAL FIRM’S CHOICE OF
ENTRY MODE
By

Bong Geul Chun

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Economics

2006

ABSTRACT

A THEORETICAL AND EMPIRICAL INVESTIGATION
OF THE MULTINATIONAL FIRM’S CHOICE OF ENTRY MODE

By

Bong Geul Chun

This dissertation provides a theoretical and empirical investigation of the
multinational ﬁrm’s choice of entry mode. In chapter I , we build a theoretical model
that shows that a joint venture is an important organizational form for a multinational
ﬁrm, although the multinational ﬁrm cannot make ex ante complete contracts with its
intennediate-input supplier. We show that if the supplier’s intermediate input is
important, then the multinational ﬁrm holds a smaller share of supplier ﬁrm’s equity to
give the supplier an incentive to produce more intermediate input. However, if the
multinational ﬁrm makes an important intermediate input, then the ﬁrm tends to hold a
greater equity share. That is, depending on the relative importance of the intermediate
inputs, the multinational ﬁrm chooses its share of the supplier ﬁrm’s equity. To test the
predictions in the theory, we use a unique data set of a newly industrialized country’s
(South Korea) multinational ﬁrms. The empirical tests partially support the theory. The
empirics show that if an affiliate has the higher ratio of intangible assets to sales, then
its parent ﬁrm is inclined to hold a smaller equity share of supplier’s ﬁnn.

Chapter H studies how the differences in a host country’s intellectual property
rights protection affect the multinational ﬁrm’s decision concerning ownership structure.
Using another South Korean ﬁrm-level data set, we ﬁnd that there is a strong negative
relationship between host country’s standards of intellectual property rights protection

and multinational ﬁrm’s equity participation in its afﬁliate. The evidence is found by

using two different samples: (1) a sample of all overseas affiliates, and (2) a sample
including only partially-owned affiliates. In addition, we ﬁnd that a multinational ﬁrm
prefers joint ownership when it invests in the resources-based sector. However, we
cannot ﬁnd any strong evidence that a multinational ﬁrm lowers its equity share of
overseas afﬁliate when the host country market is more competitive.

Chapter III examines the effects of contract enforceability and market structures on the
multinational ﬁrm’s choice between licensing and foreign direct investment. Apparently,
the multinational ﬁrm’s choice affects the host country’s social welfare. Thus, the host
country is likely to use various policy instruments such as direct restrictions, direct
subsidies, and contract enforceability to induce the multinational ﬁrm to choose a
desirable entry mode. The model shows that the host country’s preference for the
multinational ﬁrm’s entry mode crucially depends on the host country’s market
structures and the country’s capability to absorb the ﬁrm’s advanced technology. Under
certain conditions, direct subsidies to encourage foreign direct investment can beneﬁt
both the host country and the multinational ﬁrm. One interesting result is that the host
country never chooses perfect contract enforceability when the country has one

incumbent company that can compete with the multinational ﬁrm.

Dedicated to my wife and my children
For their Love, Support, and Sacriﬁces

ACKNOWLEGEMENTS

This dissertation is completed after invaluable help from many people. I would like
to begin by gratefully thanking my advisor and mentor, Professor Steven J. Matusz, who
sacriﬁced countless hours reading and correcting draﬁs, and carefully guided and
inspired me through each step of my research. I always think myself as very fortunate to
have been under his instruction. Without his time consuming guidance and
encouragement, this dissertation could never have been completed. I will remain deeply

indebted to him throughout the remainder of my career.

I am also especially indebted to Professor Susan Chun Zhu. Without her invaluable
help and encouragement, it would not have been possible for me to implement the
empirical study. She read the drafts and checked every detail of the empirical
methodology, and suggested wonderful solutions to the problems identiﬁed. In addition,
the main ideas of Chapter II were developed during discussions with her. I would also
like to gratefully thank Professors Jay Pil Choi, Carl Davidson, and Jun-Koo Kang for

their help and thoughtful suggestions.

I would like to express my appreciation to Junghwa Seo and James Kim at the
Export-Import Bank of Korea, who helped to reﬁne the data sets used in the empirical
study. I am also grateful to many colleagues and friends at Michigan State University.
Michelle Kristine Gleason and Byung-Cheol Kim deserve my thanks for having read
and commented on many parts of this dissertation. I am very thankful to Kwanghyun
Lee, who kindly shared with me his knowledge about STATA. Although not mentioned

in the list, there are still many people who helped me to go through all the difﬁculties I

faced during my stay here. I am especially pleased when I remember those days of
studying together at the library, listening to the brilliant professors, and enjoying a life at

East Lansing with friendly neighbors.

I would also like to give my special thanks to my mother and mother-in-law for their
support and encouragement. As always, I owe the greatest debt to my wife, I in Hee Bae,
for her love, support, patience, and endless sacriﬁces throughout the process of my
graduate studies. Finally, I would like to thank my children, Ye J i Chun and In Whan

Chun for the full of joy they brought me.

vi

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
CHAPTER I
OUTSOURCING, JOINT VENTURES, AND WHOLE OWNERSHIP
1. Introduction ........................................................................ 1
2. Basic Model and Results ............................................................ 5
2.1 Consumer’s Utility .......................................................... 6
2.2 Production .................................................................... 7
2.3 Incomplete Contracts ........................................................... 9
2.4 Game Stages .................................................................... 11
2.5 Bargaining over and Splitting the Revenues ............................... 12
2.6 The Optimal Equity Share ................................................... 17
2.7 Production Market Competition .......................................... 23
3. Econometric Tests ................................................................. 24
3.1 Speciﬁcation ................................................................ 24
3.2 Data .......................................................................... 26
3.3 Results ....................................................................... 31
4. Conclusion and Further Studies ................................................ 37
APPENDIX ......................................................................... 39
REFERENCES ........................................................................ 46
CHAPTER II

FIRM’S OWNERSHIP STRUCTURES: INTELLECTUAL PROPERTY
PROTECTION, RESOURCES-BASED SECTOR, AND MARKET
COMPETITION

1. Introduction ..................................................................... 49
2. Related Literature and Hypotheses .......................................... 52
3. Data and Statistical Model ................................................... 56
4. Empirical Findings ........................................................... 66

4.1 Intellectual Property Rights and Equity Participation ............. 67

vii

4.2 Quadratic Relationship and Investor Protection ........................ 73

4.3 Resources-Based Sector and Preference for Joint Ownership ......... 79

4.4 Market Competition and Equity Structures ............................ 84

5. Concluding Remarks ........................................................... 86

APPENDIX ......................................................................... 88

REFERENCES ........................................................................ 99
CHAPTER III

LICENSING, FOREIGN DIRECT INVESTMENT, AND GOVERNMENT
POLICIES

1 . Introduction ...................................................................... 102

2. The Basic Model ................................................................ 107

3. Case 1: Without Competent Local Company ................................ 112

4. Case 2: One Competent Local Company .................................... 126

4.1 Foreign Direct Investment ............................................... 128

4.2 Contract with a New Local Company ................................. 131

4.3 Contract with the Incumbent Local Company ........................ 136

5. Case 3: Two Competent Local Companies .................................. 142

6. Conclusion ....................................................................... 148
APPENDIX

1. Without Competent Local Company ..................................... 151

2. Comparison of Social Welfares .......................................... 152

3. Comparison of the Incumbent Company of Variety i’s Proﬁts . 153

4. Multinational Firm’s Choice of Entry Mode ....................................... 154

5. Comparison of Social Welfares ........................................... 155

6. Derivation of h = O ......................................................... 156

7. Comparison of Social Welfares .......................................... 156

REFERENCES .......................................................................... 158

viii

LIST OF TABLES

Table 1.1: Affiliates by Industry ...................................................... 27
Table 1.2: Affiliates by Location ..................................................... 27
Table 1.3: Year of Establishment and Affiliates by Equity Share ................ 29
Table 1.4: Afﬁliates by Income Level of Host Country .......................... 30
Table 1.5: Countries by Legal Origin ............................................... 31
Table 1.6: Equity Ownership Structures with Location Dummy ................ 32
Table 1.7: Equity Ownership Structures (above $10,000 income) ................ 34
Table 1.8: Equity Ownership Structures (above $20,000 income) ................ 36
Table 1.9: List of Industries ........................................................... 39
Table 1.10: Countries by Location ................................................... 40
Table 1.11: Descriptive Statistics .................................................... 41
Table 1.12: Equity Ownership Structures with Legal Origin Dummy ......... 42
Table 1.13: Countries by Income Level ............................................. 43
Table 1.14: Equity Ownership Structures with Income Level Dummy 44
Table 2.1: Locations of Afﬁliates and Equity Shares (%) ........................ 57
Table 2.2: Industries of Affiliates and Equity Shares (%) ........................ 58
Table 2.3: Year of Establishment and Equity Shares (%) of Affiliates 59
Table 2.4: Year of Establishment and Equity Shares (%) of Affiliates (Pure Joint

Ventures) .................................................................. 60
Table 2.5: Descriptive Statistics for Intellectual Property Rights ............... 64

Table 2.6: Descriptive Statistics for Intellectual Property Rights (Pure Joint
Ventures) ................................................................. 65

Table 2.7: Descriptive Statistics for Market Competition (All Afﬁliates) 66
Table 2.8: Intellectual Property Rights and Equity Ownership Structures ....... 69

Table 2.9: Intellectual Property Rights and Equity Ownership Structures (Pure

Joint Ventures) .......................................................... 72
Table 2.10: Quadratic Relationship .................................................. 74
Table 2.11: Countries by Legal Origin .............................................. 76
Table 2.12: Legal Origins and Equity Ownership Structures .................... 78
Table 2.13: Resources-Based Sector ................................................ 80
Table 2.14: Preference for Joint Ownership .......................................... 83

Table 2.15: Market Competition and Equity Ownership (All Affiliates) 85

Table 2.16: List of Industries ......................................................... 88
Table 2.17: Countries by Income Level ............................................. 89
Table 2.18: Locations of Afﬁliates and Equity Shares (%) (Pure Joint Ventures)

............................................................................... 90
Table 2.19: Industries of Affiliates and Equity Shares (%) (Pure Joint Ventures)

............................................................................... 91
Table 2.20: Intellectual Property Rights and Equity Ownership Structures (Pure

Joint Ventures) .......................................................... 92
Table 2.21: Affiliates in the OECD and Equity Shares (%) ...................... 93
Table 2.22: Affiliates in the US. and Equity Shares (%) ........................ 94

Table 2.23: Descriptive Statistics for Market Competition (Affiliates in the
OECD) .................................................................. 95

Table 2.24: Descriptive Statistics for Market Competition (Afﬁliates in the US.)
............................................................................... 95

Table 2.25: Market Competition and Equity Ownership (All Afﬁliates) 96

Table 2.26: Market Competition and Equity Ownership (Afﬁliates in the OECD)
............................................................................... 97

Table 2.27: Market Competition and Equity Ownership (Affiliates in the US.) 98
Table 3.1: The Values of f .......................................................... 153
Table 3.2: The Values of g ........................................................... 154

LIST OF FIGURES

Figure 1.1: Optimal and Actual Outputs of Intermediate Inputs ................ 20
Figure 1.2: Production Process .......................................................... 45
Figure 1.3: Density of Overseas Affiliates ......................................... 45
Figure 3.1: The Multinational Firm’s Proﬁts under Two Entry Modes ......... 116
Figure 3.2: Region of Parameters 6 and y for the Host Country’s Preference
............................................................................. 121

Figure 3.3: Region of Parameters 6 and y for the Host Country’s Choice of 6
............................................................................... 141

Figure 3.4: Maximum Point of Social Welfare under y = 1.3 ................... 151

Figure 3.5: Host Country’s Welfare and Multinational Finn’s Preference 151

Figure 3.6: Function f (y, 9) for Comparison of Social Welfares ............... 152
Figure 3.7: Function g(y,9) for Comparison of Proﬁts ......................... 154
Figure 3.8: The Function of v(y) ................................................... 155
Figure 3.9: The Function of b(y) ................................................... 156
Figure 3.10: The Comparison of Social Welfares ................................ 157

xi

CHAPTER I

Outsourcing, Joint Ventures, and Whole Ownership

1. Introduction

When a ﬁrm decides to purchase intermediate inputs from foreign countries, the
appropriate ownership of its intermediate-input supplier is a pressing practical question
for that ﬁrm and an important issue in economic theory. A multinational ﬁrml tries to
reap beneﬁts from technique differences and wage differences between the country
where it is located and the foreign countries where its interrnediate-input supplier is
located. In 2000, US. trade in goods involving U.S. parent ﬁrms and their foreign
affiliates - multinational ﬁrms associated with trade2 - accounted for 56 percent of US.
exports of goods and for 35 percent of US. imports of goods (Mataloni, 2002, p. 115).

Economists have made many endeavors in order to explain this huge share of intra-
ﬁrm trade. One of the important outcomes is a theory which concerns ﬁrms’ boundaries.
This theory demonstrates that the ﬁrm decides whether to produce inside its boundaries
(self-production or integration of other ﬁrms), or outside its boundaries (outsourcing),
after considering all possible costs and beneﬁts accrued from its various boundaries.
Even though the ﬁrm decides to produce inside its boundaries by integrating other ﬁrms,

dividing the equity share of the supplier’s ﬁrm between the ﬁrm and the supplier

 

I Multinational ﬁrms (MNFs) are sometimes called multinational enterprises (MNEs), transnational
enterprises (TNEs), or multinational companies, which usually have affiliates in foreign countries. We use
the term multinational ﬁrms.

2 US. multinational ﬁrms associated US. trade in goods may be disaggregated into two broad categories:
(I) Intra-MNFs trade (trade between US. parent ﬁrms and their foreign affiliates), and (2) MNF trade
with others (trade between US. parents and foreigners other than their foreign afﬁliates, and trade
between foreign afﬁliates and US. entities other than their parent ﬁrms). In 2000, the intra—MNFs exports
accounted for 21 percent of US. exports of goods and the intra-MNFs imports of goods accounted for 15
percent of US. imports of goods (Mataloni, 2002).

remains a central issue (degree of integration). Intra-ﬁrm trade is caused when
multinational ﬁrms decide to produce inside their own boundaries but outside their
national boundaries.

Since Grossman and Hart (1986) systematically analyzed the role of ex ante
incomplete contracts on the allocation of control rights (ownership), the vertical
integration of firms has been actively examined based on their property rights theory.3
However, the papers which have developed the theory on the ﬁrms’ boundaries based on
ex ante incomplete contracts have not properly focused on one very important foreign
direct investment pattern: a joint venture (or partial ownership).4 Those papers resulted
in the equilibrium of whole ownership of suppliers, pure outsourcing (no ownership), or
the coexistence of whole ownership and pure outsourcing.

Under the situation of ex ante incomplete contracts, we show theoretically that a
joint venture is one of a multinational ﬁrm’s choices. In contrast to other papers that
concern the roles of ex ante incomplete contracts, we demonstrate that a multinational
ﬁrm would choose an intermediate equity share (a joint venture) to give its
intermediate-input supplier an incentive to produce more intermediate input when the
supplier’s intermediate input is important in making a ﬁnal product. In addition, we use
a unique data set of a newly industrialized country’s (South Korea) multinational ﬁrms
to test the predictions of the theory.

There are many costs and beneﬁts involved with joint ventures.5 However, in order

to focus on the roles of ex ante incomplete contracts in the ﬁrms’ boundaries and to test

 

3 According to Whinston (2003), there are two types of theory to explain ﬁrm’s boundaries based on the
fact that the contracts are incomplete: the transaction cost economics theory and the property rights theory.
Whinston maintained that the property rights theory is more systematieal and formal than the transaction
cost economics theory.

4 Svejnar and Smith (1984) emphasized that the proportion of joint ventures of US. overseas
manufacturing operations has been growing. However, Desai et al. (2004) said that partial ownership of
foreign affiliates by US. multinational ﬁrms has declined over the last 20 years. They explained that
regulatory changes, tax changes, and reduced costs of coordinating global operations are the main reasons
for the decline of joint ventures.

5 For details of the beneﬁts and costs of joint ventures, refer to Hennart (1991).

2

them empirically, it is assumed that the goal of the firm and that of its intermediate-
input supplier are congruent when they form a joint venture.6 We assume that there are
two countries in the world, a source country (where ﬁnal-product producers are located)
and a host country (where suppliers of intermediate inputs are located), and that there
are n differentiated ﬁnal products. To produce a ﬁnal product, it is assumed that two
different intermediate inputs (x5, x” ) are needed.

Between the two intermediate inputs (x3, x” ), one of the intermediate inputs (19,)
is assumed to be produced by intermediate-input supplier in the host country,7 and the
other input (xs) is made by a ﬁnal-product producer. Thus, the ﬁnal-product producer
has to acquire the intermediate input (xH) from the intermediate-input supplier. To
produce the intermediate inputs, both producer and supplier must use their speciﬁc
factors which include physical factors such as labor, and non-physical factors such as
knowledge and effort. We assume that the quality of intermediate inputs is non-
contractible ex ante.

We show that if the intermediate input made by the producer is more important in
production of the ﬁnal product than the input made by supplier, then the producer is
inclined to hold a greater equity share of the supplier’s ﬁrm to take a larger fraction of
total revenues. Conversely, if the intermediate input produced by the supplier is more
important, then the producer is willing to lower its equity participation in the supplier’s

ﬁrm to induce the supplier to produce more of the intermediate input. That is, the

 

6 A ﬁnal-product producer represents a multinational ﬁrm (or parent ﬁrm) which has overseas affiliate (3)
although we assume that the producer makes another intermediate input, since we analyze vertical
relationship instead of horizontal relationship. In addition, the affiliates in the foreign countries denote the
intermediate-input suppliers.

7 We focuse on a multinational ﬁrm‘s entry mode into foreign countries, among them outsourcing, joint
ventures and whole ownership, under the condition that the multinational ﬁrm has already decided to
acquire its intermediate input from a foreign country. To analyze the ﬁrm’s choice of acquiring the
intermediate input from either the home country or foreign countries, other factors such as plant-level
ﬁxed costs, ﬁrrn-level ﬁxed costs, and wage differences, should be considered. Antras and Helpman
(2004) studied the determinants of the ﬁrm‘s organizational mode, such as outsourcing from domestic
ﬁrms, integration of domestic ﬁrms, outsourcing from foreign ﬁrms, or integration of foreign ﬁrms.
However, they did not deal with joint ventures.

relative importance between the two intermediate inputs affects the producer’s equity
participation in the supplier’s ﬁrm.

We assume that one party’s equity ownership of supplier’s ﬁrm is positively related
to bargaining power of that party. With a simple model, we show that the producer’s
choice of organizational mode crucially depends on the relative importance of the
supplier’s intermediate input to that of producer’s. The empirical tests with ﬁrrn-level
data partially support the prediction of the theory.

There are many papers which show the importance of joint ventures. Joint ventures
with partners located in other countries may be seen as a way of bridging cultural gaps,
while joint ventures with local investors promise to reduce political complications
(Gatignon and Anderson, 1988). Aghion and Tirole (1994) showed that when assets are
jointly used, joint ownership can be an efficient arrangement. That is to say, split
property rights can encourage innovation. They ascertained that the relative importance
between capital inputs and intellectual inputs determine the organizational aspects of
research and development (R&D) activities. However, contrary to this where a joint
venture occurs continuously, a joint venture occurred at a speciﬁc point in their model.

In environments with incomplete information, joint ownership may result in
efficient resource allocation (Crampton etal., 1987). Svejnar and Smith (1984) analyzed
important economic issues, such as resource allocation and proﬁt distribution under
various institutional scenarios, which arise in the context of joint venture operations in
less developed countries. They emphasized the role of bargaining power in distributing
the joint venture’s proﬁts. In their paper, it was implied that the relative importance of
non-contractible factors between the producer and the supplier might affect the
bargaining power of each party. We are taking this implication and formally modeling it.

The remainder is organized as follows. In section 2, we set up the basic model with

two countries, in which ﬁnal-product producers are located in the source country and

4

intermediate-input suppliers are in the host country. The model is simpliﬁed to examine
the effects of non-contractibility and relative importance of the intermediate inputs
produced by both parties on the ﬁrms’ boundaries. In section 3, we present an
econometric speciﬁcation and run regressions to test the theory empirically. Section 4

concludes and suggests some further studies.

2. Basic Model and Results

In this section, we describe a simple model and its results. The model is similar to
Antras and Helpman (2004), and Antras (2003) models. However, the implications are
very different from those of the two studies. Antras and Helpman (2004) have adopted a
typical North-South model: different productivity levels and a different wage. They
have explained an equilibrium in which ﬁrms with different productivity levels choose
different ownership structures and suppliers’ locations, such as foreign direct investment
in the North or South, or outsourcing from the North or South.

Antras (2003) focused on two different patterns in international trade, one is that the
share of intra-ﬁrm imports in total US. imports is positively related to the capital
intensity of the exporting industry, and the other is that the share of intra-ﬁrm imports of
the total US. imports is positively related to the capital-labor ratio of the exporting
countries. He analyzed these patterns theoretically and tested them empirically.

However, Antras and Helpman, and Antras, assumed that the bargaining power of
each party (either the parent ﬁrm or the supplier) is ﬁxed and exogenously given. As a
result, they did not consider one of the ﬁrm’s most important organizational modes:
joint ventures, or intermediate equity participation. In addition, neither of these papers
tested their theories using ﬁrm-level data. We show a joint venture is one of the ﬁrm’s

optimal organizational choices, and we test the theory with ﬁrm-level data.

It is assumed that there are two countries, a source country where the producer (the
headquarters of multinational ﬁrm) of variety 1' is trying to ﬁnd a partner in a host
country where many intermediate-input suppliers want to be the producer’s partner.
Thus, the producer can choose a share of supplier ﬁrm’s equity. In this world with two
countries, there are n differentiated ﬁnal products. It is assumed that there are no
transaction costs and no differences in scale economics between the plant-level and the

ﬁrm-level.

2.1 Consumer’s Utility
In this world, the identical utility function of consumers is represented by
.1.
U=[J;'y(r)adi]" (1.1)
where 0 < a <1, and n is the number (measure) of varieties which is assumed to be
very large.8 In this consumer’s preference, the elasticity of substitution between two

varieties is I/(l —a). Income is deﬁned as Y a gp(i)- y(i)di . From the consumer’s

preference, the demand ﬁrnction for variety i is given by

_l_
y(i):_p(_i_):-I__.Y (1.2)

a

p(j)‘-’T‘dj
_a_
where Y is income. If we deﬁne gp(j)a‘ldj as the price index, P,9 then the

inverse demand function for variety i can be described simply.

p( i) = y(Oar—1 _ Yl—a . Pa-l

 

' Does the market competition really affect the ﬁrm’s choice of entry mode or not? In order to address the
implications of this question, this consumer utility function is introduced.
9 It is assumed to be ﬁnite and constant.

where Y2 Kp(i).y(i)di and PE Ep(j)5q‘_‘dj.

2.2 Production

In order to produce one unit of y , the ﬁnal-product producer needs two different
intermediate inputs, xs and x” , where subscripts S and H represent the source
country and the host country, respectively. The intermediate inputs x8 and x” are
produced by the, ﬁnal-product producer in the source country and by the intermediate-
input supplier in the host country, respectively. However, we assume that the producer
does not require costs to produce the ﬁnal product with the intermediate inputs.

To produce the intermediate input x5 , the ﬁnal-product producer needs speciﬁc
factors, which include physical factors such as labor, as well as non-physical factors
such as knowledge and effort. The intermediate-input supplier uses speciﬁc factors to
produce 27,, as well. By extension, x3 can be interpreted as being produced by using
relatively abundant knowledge assets in the source country, and x” by using relatively
abundant knowledge assets in the host country.

We can conclude that the ﬁnal product y is made through multi-tiered production
stages. First, different speciﬁc factors are needed to produce the intermediate inputs xS
and x” , respectively. Second, these intermediate inputs, xs and x” , are used to
produce the ﬁnal product y . The producer in the source country has to choose a
partner (the supplier of x”) in the host country to be able to make the ﬁnal product y .
Figure 1.2 in Appendix shows the production process for the ﬁnal product y .

It is assumed that the production ﬁrnction for the ﬁnal product y is in Cobb-

Douglas form, such as

]_
y=(xs)¢(x.) ¢ (13)
where ¢ and 1—¢ represent the relative intensity and importance of its and x” in

7

the ﬁnal product’s ( y) production, respectively. We can approach this slightly
differently. Following Svejnar and Smith’s paper, ¢ and 1—¢ might be related to the

bargaining power of the producer and the supplier, respectively.10 In order to produce

the ﬁnal product y , if xs is more important than x” (showed by a larger ¢), then

the producer has more bargaining power than the supplier.

We assume that the quality of the intermediate inputs xs and x” should be above

a certain level to produce y. The quality problem is discussed in the following section.
In order for each ﬁrm to produce high quality intermediate input (xi where .i = S, H ),
the producer / the supplier must pay costs. The production cost function is assumed to
be in quadratic form.

1 2 .
Ci=§7ixi where z=S,H and yi>0 (1.4)
This has the usual characteristics of common cost functions, such as the ﬁrst (C15(xi))
and the second (C;(xi)) derivatives being positive, and thus the curve for cost function

is convex. In this cost ﬁrnction, 7i is inversely related to the country i’s cost

effectiveness in producing the intermediate input.

In this world with two countries, n differentiated ﬁnal products are made in the
source country. It is assumed that there is only one producer in variety 1' (where
i=1,2,...,n), and many suppliers are competing to be a partner of the producer. The
producer of each variety has limited monopoly power because of the many

differentiated ﬁnal products.

 

'0 This statement is true because of the assumption that one party’s bargaining power is positively related
to how much equity share of supplier’s ﬁrm the party holds.

2.3 Incomplete Contracts

Recall that the speciﬁc factors of the producer and the supplier are needed to
produce the intermediate inputs xs and x” , respectively. However, these intermediate
inputs are used to produce y only if the quality of these intermediate inputs is above a
certain level. If the quality of the intermediate inputs is under said level, then the inputs
are assumed to be useless. We also assume that the costs to produce the lower quality
level of the intermediate inputs are negligible.

An important consideration is that the quality of the intermediate inputs is assumed
to be non-contractible ex ante.ll That is, a third party cannot verify the exact quality of
the intermediate inputs. Under the ex ante incomplete contracts situation, if the producer
contracts to buy a certain type of x” for a certain price, then the supplier has an
incentive to produce low quality intermediate input at a lower cost and still receive the
contracted price since a third party cannot verify the quality of the input.

It is a crucial assumption that the quality of intermediate inputs is observed by two
parties (the producer and the supplier), and that an outside party such as a court of law
cannot verify the quality. This ex ante incomplete contractibility is identical to the fact
that the outside party cannot verify the production costs of the intermediate input. Like
other papers such as Antras (2003), Antras and Helpman (2004), and Grossman and
Helpman (2002), the quality problem is dealt with through the quantity problem.

The problem is that the quality of intermediate input is always lower than the
socially optimal level because of ex ante non-contractibility. The quality problem is
considered being identical to the quantity problem in that the quantity of the

intermediate inputs is also produced at a lower level than the socially optimal level. The

 

” The quality of intermediate input could be considered partially non-contractible ex ante as Antras
(2003) mentioned. However, he asserted that partially non-contractible contracts do not change the results
under certain conditions. We thank Antras for sharing the proof of his assertion.

9

producer / the supplier must use more of its own speciﬁc factors to produce a higher
quality (or larger quantity) of its intermediate input but bear all costs. The producer
cannot make ex ante enforceable contracts specifying the purchase of the intermediate
input x” for a certain price. For the same reason, the producer cannot sign ex ante
enforceable contracts with its suppliers specifying the division of total revenues.
Between the two parties, only the allocation of residual rights is contractible ex ante. '2

It is also assumed that the producer and the supplier have to specialize their
intermediate inputs for their ﬁnal product, y. With ex ante non-contractibility, the
specialization gives rise to a hold-up problem. In other words, the intermediate inputs
are useful only for the speciﬁc ﬁnal product13 y , and cannot be used to produce other
ﬁnal products (a relationship-speciﬁc problem). Because of incomplete contracts and
the hold-up problem, the two parties bargain over total revenues. ‘

After producing its intermediate input, each party can threaten to refuse delivery of
its own intermediate input. The production costs of intermediate inputs are bygone
(sunk cost) when they bargain over revenues ex post. The fraction of total revenues
which the producer / the supplier can take depends on how much bargaining power the
producer / the supplier has. It is assumed that the bargaining is costless; as a result,
costless bargaining yields an efficient ex post outcome. The bargaining power of one
party is assumed to be positively related to the share of supplier ﬁrm’s equity which that

party holds.

We show that the two parties do not choose a socially optimal output level of the

 

'2 Grossman and Hart (1986), and Hart (1995) argued that contractual rights can be divided into two
types: speciﬁc rights and residual rights. When it is too costly for one party to specify a long list of the
particular rights it desires over another party’s assets, it may be optimal for that party to purchase all the
rights (residual rights) except those speciﬁcally mentioned in the contract. They emphasized that
ownership is the purchase of these residual rights of control. That is, the owner of an asset has residual
rights over that asset: the right to decide all usages of the asset in any way not inconsistent with a prior
contract, custom, or law.

'3 This assumption could be relaxezd from perfect specialization to partial specialization. 'lhe supplier
can choose the degree of speciﬁcity of its intermediate input, considering the value within the relationship
and value outside. As a result, the degree of speciﬁcity can be determined endogenously. For more detail,
refer to Grossman and Helpman (2002).

10

intermediate inputs since they are not compensated fully. The share of equity which the
producer holds decides the producer’s fraction of the total revenues made by selling the
ﬁnal product, y. Grossman and Hart described that “through its inﬂuence on the
distribution of ex post surplus, ownership rights will affect ex ante production
[investment] decisions. That is, although ex post efﬁciency is guaranteed under any
ownership structure, [each] ownership structure will lead to a different distortion in ex

ante production [investment] (Grossman and Hart, 1986, pp 696-697).”

2.4 Game Stages

The game in this model consists of four stages. At the ﬁrst stage, a producer in the
source country chooses a supplier in the host country and decides the most efficient
entry mode. In other words, the producer chooses a share of I supplier ﬁrm’s equity
(OS 3 S], where 3 represents a share of supplier ﬁrm’s equity which the producer
holds). When the entry mode is decided, the allocation of residual rights is made. At the
second stage, the producer and the supplier produce the intermediate inputs xs and
Jr” , respectively, by using their own speciﬁc factors. At the third stage, as the quantity /
quality of the intermediate inputs is validated by the two parties, bargaining between the
two parties is carried out. If they arrive at an agreement, the ﬁnal product is produced by
using these two intermediate inputs, x5 and x”, in the next stage. An agreement
results in dividing the fractions of total revenues the producer and the supplier can take.
The next subsection explains the bargaining processes. Finally, the producer makes the
final product without incurring any costs, and sells the product in the market. The fourth

stage does not play any role.

2.5 Bargaining over and Splitting the Revenues

Since the intermediate inputs are assumed to be relationship-speciﬁc, neither party
can sell the inputs xs and x” to outside parties. However, the quality of the inputs
should be above a certain level to produce the ﬁnal product. After bearing their own
sunk costs for the production of intermediate inputs, both parties bargain over total
revenues. Based on the consumers’ preference, the demand function for the ﬁnal
product is derived as equation (1.2). Thus, the total revenues are represented as

R = p-y = Yl-aPa-lya : Yl-aPa-1(xs)a¢(xﬂ )a(1-¢) (1.5)

There are many factors which affect the bargaining power of each party. Bargaining
power may determine the distribution of beneﬁts / revenues between the producer and
the supplier. Bargaining power might enable one party to gain control of a joint venture
which is likely to manifest itself in a larger equity share. Behrrnan and Grosse (1990)
explained that bargaining theory can be applied to both parties’ relations, focusing on
the key strengths of producer and supplier that enable each to obtain more favorable
results.

For example, if a local supplier has comparatively advanced technologies, then the
supplier can be expected to have the upper hand in bargaining and to obtain favorable
results, such as a high degree of local ownership and control. Svejnar and Smith (1984)
showed that the bargaining power determines the division of a joint venture’s proﬁt.
Fagre and Wells (1982) emphasized that equity ownership is seen as an outcome of
negotiation: a representation of relative power between two parties. They found that the
degree of multinational ﬁrm’s ownership of Latin American affiliates was based on the
ﬁrm’s technology and advertising intensity.

Lecraw (1984) also found that the level of equity participation of multinational
ﬁrms is inﬂuenced by their bargaining position. He showed the greater ability of

12

multinational ﬁrms to obtain higher equity ownership and to acquire more control over
affiliates as the ﬁrm’s technology and advertising intensity increased.l4 Kobrin (1987)
studied the bargaining model using data from 563 affiliates of US. manufacturing ﬁrms
in forty-nine developing countries. He explained that bargaining power is a function of
either ﬁrm-speciﬁc assets or ownership advantages and imperfect markets that allow the
ﬁrm to contain the advantages. He demonstrated that multinational ﬁrms have
advantages or ﬁrm-speciﬁc assets, and the ﬁrm’s possession of those results in a ﬁrm’s
preference for ordering economic transactions through internal administrative
hierarchies (vertical or horizontal integration), rather than the external market (p. 609).
In addition, he demonstrated by giving a few examples that ﬁrm-speciﬁc assets can be
exploited more efﬁciently through internalization with deﬁnite control. He found that
the degree of a multinational ﬁrm’s ownership of affiliates was a function of bargaining
strengths.

Based on their ﬁndings, it is assumed that the producer’s bargaining power is
positively related to its share of the supplier ﬁrm’s equity. On the contrary, the
bargaining power of the supplier is assumed to be negatively related to the producer’s
equity ownership. The producer can give some bargaining power to its supplier in order

for the supplier to use more speciﬁc factors and to produce more x”. From the
bargaining, it is assumed that the producer receives a fraction n(s)e (0,1) of the ex

post gains from the relationship and this fraction increases with respect to equity share

which the producer holds (77'(s)>0, where s represents a share of supplier ﬁrm’s

equity which the producer has).

 

'4 A multinational ﬁrm is inclined to have a greater equity share if the advanced technology of the
multinational ﬁrm can be spilled over to its affiliates or other local companies in the host country. For
example, only the producer has the full knowledge to assemble the intermediate inputs to make a ﬁnal
product. Continuous contacts between the producer and its supplier can make it easy for the supplier to
imitate the producer’s advanced technology to assemble. Thus, the producer tends to hold a greater equity
share to control its supplier and prevent it from imitating the technology. The issues on technology
spillovers are dealt with explicitly in the following chapter.

13

The outside option15 for the producer is different depending on how much share of

supplier ﬁrm’s equity the producer holds. In cases where the producer has an equity

share majority (0.5 < s s1), it is assumed that the producer has the residual rights.16 If

the producer and the supplier cannot arrive at an agreement, the producer has the control

rights on the fraction, 6 , of x” . because the supplier’s ﬁrm is under the control of the
producer. Thus, the producer can directly order to pass over the fraction 6 of x”.

This fraction 6 is independent of the equity share when the producer has an equity
share majority (0.5 < s s 1) because the intermediate input x” is of no use to outside
producers. That is, x” is useless for the supplier if the two parties walk away from
their deal. Since the outside option for the supplier is still zero no matter how much of

the equity share the supplier has under the situation when producer has an equity share

majority ( 0 3 s3 < 0.5 , where ss is an equity share which the supplier has
(ss =(1-s))), the producer tries to acquire the maximum 6 .

On the contrary, when the supplier has 50% or more of the equity share
(0.5 s ss 3 l ), it is assumed that the supplier has the residual rights. In this case, the
producer cannot obtain anything if they do not reach an agreement.l7 The ﬁaction 6
(default ﬁ'action) should be strictly less than one. If 6 is equal to 1, then the supplier
receives nothing ex post but only bears sunk costs to produce the intermediate input x”
when the producer has an equity share majority. As a result, the supplier’s optimal
production of intermediate input is 0 when the producer can control the supplier. That is,

if 6 =1, then the majority ownership (0.5 < s S 1) never occurres. Thus, the 6 should

 

'5 The actions chosen when the two parties leave the negotiating table.

'6 The conﬂicting policies between the producer and the supplier might not be solved easily, and the
objectives of both parties might not be congruent. This could be the reason why joint ventures have lasted
shorter than whole ownership. However, for simplicity, we assume that one party’s interest is congruent
with that of the other party.

'7 We make a trivial assumption that if the producer has 50% of supplier ﬁrm’s equity (3 =0.5), then the
producer cannot control the supplier ﬁrm.

l4

be in the range of [0,1) and is assumed to be close to 0.18

In summary, the ﬁrst case is when the producer has an equity share majority

(0.5 < S S 1). If the two parties reach an agreement, then the producer obtains the default

fraction of intermediate input, 6x” , which is identical to 6a(1_¢)R, and additional

(10

n(s)(l — 6 -¢))R through bargaining over the remaining revenues, (1— 6°1'(1_‘15))R.19

Then, the supplier can obtain (1 -r](s))(l —6a(l—¢))R by bargaining over the
remaining revenues. If the two parties do not arrive at an agreement, then the producer

can acquire only the default option, 6a(]_¢)R, and the supplier cannot receive
anything.
The second case is when the producer does not have an equity share majority

(0 s s < 0.5 ). Under the agreement, the producer can obtain r7(s)R and the supplier
can receive (1 -77(s)) R. If they do not reach an agreement, both of them obtain nothing.

As a result, reaching an agreement is always preferable (dominant equilibrium) for both
parties in the ﬁrst case as well as the second case.

Under the agreement situation, it is easily derived that the fraction of revenues
which the producer obtains is an increasing function of its equity share, 3 .20 The

differentiation of the fraction which the producer can obtain with respect to s is

 

'8 We make this assumption for simplicity. We can take 6 to be far from zero. However, the fact that the
fraction of total revenues the producer can obtain is monotonically increasing with respect to equity share
(s ) the producer holds makes a difference.

'9 This remaining part is called quasi-rent. Quasi-rent is the excess of its value over its salvage value, that
is, the value of its next best use to another renter (Klein et al., 1978).

2° The residual rights (ownership) affect the distribution of ex post revenues through its effect on the
outside option in this model. Although we assume that the two parties’ interests are congruent, we can
relax that assumption. If we assume that the residual rights are decided by an unanimous vote, then the
producer cannot have the residual rights when it partially owns the supplier ﬁrm’s equity (OS 3 <1)
because the supplier can exercise its veto. That is, the producer has the residual rights only if it wholly
owns equity share. In this case, there is no discontinuous point when s is in the range of [0, 1). However,
there is ajurnp when S = l . As a result, only the bargaining power plays an important role in splitting the
revenues under an assumption that the residual rights are decided by an unanimous vote.

15

a[a“(“¢) + ”(9(1 —5"(‘“¢) )]/as > 0, when 0.5 < s s]

6[77(s)]/as > 0 , when 0 s s s 0.5
The other interesting thing is that the ﬁaction of revenues the producer can obtain is
increasing in ¢, the relative importance of x3, given the equity share 3, when the

producer has an equity share majority.
a[5a(“¢) + r7(§)(l — 604"” )]/a¢ > 0

where 29' is the given equity share (0.5 < E s 1).
For simplicity, we deﬁne ﬂ(s) as the fraction of revenues which the producer can

obtain, over the whole range of the equity share ( s 6 [0,1]).
,6(s) a 5a(1—¢) + 17(s)(1— 6a(l-¢)) , when 0.5 < s 51
ﬂ(s) a 77(s) , when 0 s s s 0.5

We deﬁne ,B and ,6 as the ﬁaction of revenues which the producer obtains when

it outsources intermediate input (xH) through arms-length arrangement (s = 0) and

wholly owns the supplier’s ﬁrm (5 =1), respectively.
g 5 77(0)
g E 77(1)(1-5a(1—¢))+§a(1-¢)

Although the ﬁrst derivative of ﬂ with respect to s (,B'(s)) is not well deﬁned when
s is 0.5, it is not a serious problem.” In reality, most of afﬁliates (suppliers) of Korean
multinational ﬁrms (producers) are owned through a majority equity share (3 > 0.5) as

is seen at Figure 1.3 in Appendix. Furthermore, we showed that ,6 is monotonically

 

2' There is a discontinuous point when 3 =0.5. However, the discontinuity does not cause any problem
since the fraction which the producer receives is increasing in S overall, and 6 is assumed to be close
to 0. This discontinuity at S=0.5 might explain the empirical ﬁndings by Blodgett (1991) that 50-50
joint ventures are the preponderant type overall (43.1%). However, these 50-50 joint ventures do not
occur frequently in the Korean case (seven 50-50 joint ventures out of 124 joint ventures over all).

16

increasing with respect to s in the range of ,B(s) e [g E].

It is assumed that the explicit formula of ,6(s) is known. Thus, the equity share 3
can be easily obtained for a given ,6. At the ﬁrst stage, the producer decides the mode
of entry by choosing an equity share s. From now on, we focus on the optimal fraction
of revenues ([30) because solving for an optimal ,60 is much easier than ﬁnding so.
Based on the assumption that the explicit formula of ﬂ(s) is known, if the optimal
,B" is chosen, then the optimal equity share so can be obtained without any

difﬁculties.

2.6 The Optimal Equity Share

This model is solved by backwards induction, from stage four to stage one. For
simple denotation, we omit s ﬁ'om now on. Thus, ﬂ represents ,B(s). As the
producer can obtain only the fraction ,6 of revenues, it maximizes its proﬁt to choose
the intermediate input xs at the second stage. From equations (1.4) and (1.5), the

producer’s problem is

_ _ - 1
a}? ﬂYl aPa 1(xS )a¢(xH )a(l ¢) ‘57ng‘

Similarly, the supplier chooses x” to maximize its proﬁt

_ _ 1.. 1
IgZXO-ﬂﬂ’l “P“ ‘(xs>“"(x,.)"‘ ”—57“;
After solving these two equations simultaneously, the optimal x3 and x2, are

obtained by the producer and the supplier, respectively.

 

 

 

2-a(l-¢)
x3 =ﬂ 2(2—0)
l
(20 a) €1'(1‘-24¢5)-2 “(12'1” 2—a (1-6)
x(1-,6)2'(2'-a) arYl‘O’Pa"l (13—) M
¢ 7 H

 

17

1
Q aQ—Z 7—}?

_0L 2.6,
.;-,2<2—a>(1-,.)mg .Yx-aPa-IW (71%] 2 (1.7)
73 -

From these two equations, it is easily derived that the demand for the intermediate

input x3? is negatively related to 73 (8x? /6yS < 0 ). This result is reasonable because

if the production cost effectiveness for the intermediate input in the source country is

deteriorating, then the demand for the intermediate input x‘S’ is decreasing when all the

other variables are unchanged. The interesting thing is that ax. 67. (where i: j) is
I J

negative. That means the j country’s cost effectiveness is negatively related to the

demand for 1' ﬁrm’s intermediate input since an increase of y j causes the j ’5 cost

for the intermediate input to rise.
When inserting two equations (1.6) and (1.7) into equation (1.3), then the ﬁnal

product is represented by parameters, income Y , and price index P.

1
I _
y= arl-apa-I[_¢ﬂ]¢[a_-z>o_—a]“” ’ 6'

 

(1.8)
75 7 H
The total revenues and the price of the ﬁnal product are represented as
¢ (1 ¢) ‘9—
2(1—a) 2(a-l) _ _ - 2-a
1H 2-.. P 2.... (4%) [will] (,9,
73 7 H
¢ a_-1
.111 51. (Ht) 2—a
p = Yl-aPa—lya-l = y2—ap2-a a[ﬂ] [W] (1.10)
75 7 H

We can obtain the socially optimal output levels of the intermediate inputs. If we
assume that contracts are complete ex ante, then the privately chosen level is equivalent

to the socially optimal production level. The problem changes as shown

18

l—a a—l a¢ a(1—¢)__1_ 2_l 2
391%: Y P (x8) (xH) ZnyS 2nyH

When solving this equation, we can ﬁnd the socially optimal production levels of x;

t .
and Jr” , respectlvely.

1
a(1-¢)-2 a(12—Q) 7:0?

:1: 2 _
x5: aY““P““ 3,-5- M (1.11)

¢ 7 H

1
gig org-2 2:517
2
x" = aYl'aPa_l [31—] [L] (1.12

” 73 (1—¢) )

The socially optimal output of intermediate input is never achieved because of ex
ante non-contractibility and the hold-up problem. This can be proved simply. When
equations (1.6) and (1.7) are divided by equations (1.11) and (1.12), respectively, then
the value should be equal to 1 if and only if the socially optimal production level is
identical to the privately optimal output level.

—‘ _

x3

1
o 2-a<I—¢> zeta—g) 7:2? Zia—ﬂ 2a(1—_¢)
x5 {3 2 (1-5) 2 ] zp 29‘“) ~(1—ﬂ)2(2-a) <1 (¢1)

2-aQ
x”- 2 2(2‘“).(1—,6)2(2—a)<1 ($1) I

l
0 931 2-0.2 7:07 _£ﬂ_
T - 2 “(1 — .3) = ,5
x11
Thus, an intermediate input is always produced at a level lower22 than the socially
optimal production level. In other words, the ﬁrst best outcome is not available. As is
seen from Figure 1.1, actual production points of the intermediate inputs are somewhere

inside of the shaded rectangle. Although these ex ante inefficient output levels of

intermediate inputs results in ex post inefﬁciency, the producer chooses an equity share

 

22 Under the incomplete contract situation, there might exist an over-production problem as well.
Whoever has the control rights causes the over- (or under-) production. Examples are found in Tirole
(1986)

19

to maximize its proﬁt.

Figure 1.1: Optimal and Actual Outputs of Intermediate Inputs

 

xH

 

 

 

(x§,xH ) -

c
~..

 

 

 

 

 

 

The producer decides its ﬁrm’s boundaries by choosing an equity share 3. As
mentioned before, we focus on the optimal ﬂ°(s) , because if the optimal ,6” is

chosen, then the optimal equity share so can be easily obtained. The producer chooses

x3 to maximize proﬁt

1 2 l 2
ggx7r=ﬂ'R-E}’st =ﬂ'P'y‘-2'7sxs

= ﬂY 1—a Pa-lxg¢xZ(I-¢) — % 73x3

The ﬁrst order condition is
ﬂa¢Y1_aPa_1xSa¢-le(I-¢) — 78x3 = 0
That is

_ _ 1-
75x52‘ =ﬂa¢Y1 “Pa lxg¢xg( 4’)

I 2 l
:>— x =— a R
275 H 216 ¢

20

At the next stage, the producer picks ﬂ

1 2
Irf =ﬂ-R—Eysxs

a

2(l—a) 2(a—l) ¢ I—¢ m
=[1_la¢)ﬂr 2-0 P 2-0 a[¢—ﬂ-] [———“‘¢X"”)] (1.13)
2 75 7H

 

We have to maximize equation (1.13) with respect to ,6. The solution of the above

problem is identical to that of the problem below

 

a
2_a[¢lnﬂ+(l-¢)ln(l-ﬂ)]

mgx In ,8 +
The ﬁrst order condition23 is

1+_a_[g_(1—¢)]_0

ﬂ 2 -a .3 (1 - ﬂ)
When this equation is rearranged, then
ﬂ” = til—3:22 (1.14)

We can have a unique so for given a and ¢. If we differentiate (1.14) with respect

to ¢,then
2Z=£>0
(M 2
Thatis
0
ﬂ>0 => ai>0 I
6¢ 5¢

This result means that depending on the relative importance of the intermediate
inputs, the producer chooses its share of supplier ﬁrm’s equity to raise its proﬁt. If the
supplier’s intermediate input is very important, then the producer gives the supplier a
larger fraction of total revenues to induce the supplier to produce more intermediate

input. In that case, although the producer’s ﬁaction of revenues decreases as it lowers its

 

 

 

l 1 -
23 The second order condition is satisﬁed: —[—2 + a (1— ( ¢) )] < 0.

ﬂ 2—a 1:2 u-mz

21

equity ownership, the producer’s net revenues increase since the total revenues grow
more rapidly. As a result, the producer can make a greater proﬁt in spite of the smaller
fraction of revenues.

The result implies there might be some reasons the producer in the source country
holds an intermediate share of equity, although the technologies in the source country
are more advanced than those in the host country. In other words, it is possible that the
producer would choose a joint venture if the supplier makes important intermediate
input whose quality is ex ante non-contractible, although its intermediate-input supplier
is located in the less developed country. We test this result empirically in the following
section.

We can solve the socially optimal fraction of revenues which the producer receives,
[3* , by using equations (1.11) and (1.12). Given the optimal levels of x; and x; , we

should choose the optimal ,6 in order to maximize total revenues. Then, the socially

optimal fraction ,8. is equal to ¢. It is implied that the social planner must give the

producer the fraction of revenues which is exactly equal to the importance of its
intermediate input.
However, the producer has a larger fraction of revenues than the socially optimal

fraction because the producer has the right to choose an equity share. If we subtract ,6”

from ,6}. , then

ﬂo-ﬂ" = 2—a:I—¢)_¢= (Z—a;(1—¢) >0

When the production function is in Cobb-Douglas form such as y = x? -x,1f¢ , and pi

prS ____¢

(p y) is the price of xi (y , respectively), the cost minimizing solutions are
Py

22

x
and mﬂ—d It is shown that the socially optimal fractions of revenues are
Py

exactly related to the cost minimizing solutions.

2.7 Production Market Competition

If the market becomes more competitive, then the producer decreases its share of

supplier ﬁrm’s equity. In this model, the elasticity of demand between two varieties is

 

represented by and the mark-up of ﬁnal product is (i). From these formulae,
a

(l-a)
an increase in a implies that the market becomes more competitive. When we

differentiate equation (1.14) with respect to a

£42
60: 2

Since the relative importance (¢) of producer’s intermediate input is less than 1, the

< 0, where a 6 (0,1)

differentiation with respect to a should be negative.

0 0
6'8 <0 => §S—<0 I.

That is,
6a 6a

 

This result means that if the competition in the market increases, then the producer
decreases its equity share of supplier’s ﬁrm. This might be because the producer faces a
less proﬁtable market.24 This implies that if the producer loses its dominant power in
the market as the market becomes less concentrated, then the producer would hold a
smaller equity share of supplier’s ﬁrm.

Fagre and Wells (1982) found that a larger number of producers (parent ﬁrms) per
industry is associated with lower equity share of their suppliers (affiliates) relative to

other industries. Gatignon and Anderson (1988) explained,

 

2‘ Empirically, this has not been examined thoroughly. Blomstrom and Kokko (1997) emphasized in their
survey paper that “most authors have not been able to — or have not even tried to — determine whether the
high degrees of concentration in the industries where foreign affiliates are present have been caused by
multinational ﬁrms or whether multinational ﬁrms have just been attracted to these industries by good
proﬁt opportunities (p. 30).”

23

Caves (1982) has noted that non-American multinational ﬁrms appear to be willing to
employ lower-control governance structures more frequently than American
multinational ﬁrms, perhaps because non-US. multinationals often have less market
power. Hence, the US. origin of the parent corporations may account for part of the
tendency to assume higher-control entry modes. Interestingly, the business press
reports a growing willingness to substitute lower-control entry modes for complete
ownership and attributes the trend to increasingly formidable competition, thus

reducing the multinational’s freedom to exercise its preferences (p. 333).

3. Econometric Tests

In this section, the Korean ﬁrm-level data set is used to test the hypothesis25 of the
previous section. The hypothesis is that if an afﬁliate’s intermediate input is important,
then the parent ﬁrm of that afﬁliate holds a smaller share of afﬁliate’s equity.
Conversely, if the parent ﬁrm’s intermediate input is important, then it owns a greater

equity share. The ordinary least squares method is applied to test this hypothesis.

3.1 Speciﬁcation

The hypothesis being tested is that if an affiliate (supplier) produces more important

intermediate input x” than that the parent ﬁrm (producer) produces, then the parent

ﬁrm holds a smaller share of afﬁliate’s equity. However, if the parent ﬁrm makes more

important intermediate input x5 , then it holds a greater share of aﬂiliate’s equity. It is

assumed that the parent ﬁrm and its affiliate use own speciﬁc factors to make the

intermediate inputs xS and x” , respectively.

 

2‘ We do not have enough observations in the data set used in this chapter to test the relationship between
market competition and the producer’s equity participation. However, we test the relationship in the
following chapter, where we use a different data set.

24

In practice, the importance of affiliate’s (parent ﬁrm’s) intermediate input is
revealed by how much the affiliate (the parent ﬁrm, respectively) has speciﬁc factors to
make its intermediate input. As we have examined in the previous section, if a supplier
has more speciﬁc factors and uses the factors to produce a higher quality (larger
quantity) of intermediate input, then the producer holds a smaller share of supplier
ﬁrm’s equity to give the supplier a larger fraction of revenues. This incentive scheme is
used to induce the supplier to use more speciﬁc factors to produce a higher quality
(larger quantity) of intermediate input under the ex ante incomplete contracts situation.

We us the ordinary least squares method to test the hypothesis. For empirical tests,
the independent variable should be decided, which represents speciﬁc factors used to
make intermediate inputs. To measure each party’s speciﬁc factors, we use ﬁrm’s
intangible assets which are available in the data set. Kimura and Pugel (1995)
emphasized that intangible assets are considered to be a major source of competitive
advantage that can inﬂuence foreign direct investment decisions.26 The intangible assets
are deﬁned as something of value that cannot be physically touched, such as a brand,
franchise, trademark, or patent. The intangible assets usually incur non negligible sunk
costs. We use the ratio of intangible assets to sales as the independent variable.

The following speciﬁcation is made to test the hypothesis

log(equityi) = v + g] log(piti) + g2 log(sitlj ) + 01151.03 + 81-]. (1.15)
where subscript i represents a parent ﬁrm i , and log(equityi) is the logarithm value
of share (%) of afﬁliate’s equity which a parent ﬁrm 1' holds. The piti is the parent

ﬁrm’s ratio of intangible assets to sales, and sitij is the ratio of intangible assets to

 

2" They used R&D-intensity and advertising-intensity to reflect the development of technological
intangible assets and important aspects of overall marketing intangible assets, respectively. We do not
have those intensities in the data set, but parent ﬁrm’s and afﬁliate’s intangible assets are available.

25

sales of 1' parent ﬁrm’s afﬁliate j. The 1/15. is a vector control, and al.. is an error

H J

term. The theory expects a positive sign of g], and a negative sign of g2.

For the control variables, a parent ﬁrm’s size and an affiliate’s size measured by
total sales are added to control the possibility of parent ﬁrm’s resource constraints. For
example, if the parent ﬁrm cannot wholly own afﬁliate’s equity because the affiliate is
too big to acquire, then the parent ﬁrm might prefer partial ownership to whole
ownership. To control the differences in affiliate’s location and industry, dummy
variables are added. To control the possibility that the variation of laws protecting
investors matters for parent ﬁrm’s ownership patterns around the world,27 we classify
the host countries on the basis of the origin of initial laws. We also add an institutional

variable to control the effects of host country’s characteristics on equity ownership.

3.2 Data

The Korean ﬁrm-level data are found in the Korea Investors Services Financial
System and Korea Investors Services Stock Market Analysis Tool database of the Korea
Investors Services Co., Ltd, which contains the parent ﬁrms’ balance sheets and proﬁt
and loss statements of 2002. The summarizing information on the Korean multinational
ﬁrms’ overseas affiliates is published by the Export-Import Bank of Korea. Limited
information on the overseas affiliates has been received from the Export-Import Bank of
Korea.28 We use the Korean parent ﬁrms’ and their overseas affiliates’ ﬁnancial
statements of 2002.

Among 15,655 overseas affiliates in 2002, the data set includes 318 overseas

affiliates and 129 Korean parent ﬁrms of these affiliates across all industries. The

 

2’ La Porta et al. (1998).
28 We grateﬁrlly thank Junghwa Seo and James Kim of the Export-Import Bank of Korea for helping with
reﬁning the data set.

26

Export-Import Bank of Korea has the rights to collect the balance sheets and proﬁt &
loss statements of overseas affiliates annually from any Korean parent ﬁrm which has
invested over 10 million dollars to acquire a share of overseas afﬁliate’s equity. Thus,
we have information on Korean ﬁrms’ overseas afﬁliates in which over 10 million
dollars has been invested by any Korean ﬁrm. Although we have a sample selection
based on the investment amount, it does not cause a serious problem because we still
can have unbiased and consistent estimators of the parameters. In the data set, the ﬁrst
overseas affiliate of a Korean ﬁrm started to run a business in 1973. However, the

majority of the overseas affiliates have engaged in business since the 19905 (Table 1.3).

Table 1.1: Afﬁliates by Industry

 

No. of Afﬁl. Average”(%) Median”(%) St.dev. ') Sales ($ mil.)

 

 

Manufacturing 166 88.0 100 1 7.8 27,564
Non-manufac. 1 52 87.8 100 24.5 55,667
Whole industry 318 87.9 100 21.2 83,231

 

Note: 1) equity shares.

Table 1.2: Afﬁliates by Location

 

No. of Afﬁl. Average”(%) Median”(%) St. dev.” Sales (3 mil.)

 

 

Asia 167 76.4 96.0 22.2 32,401
North America 61 98.8 100 16.2 31,264
Europe 59 55.5 100 21.5 17,112
South America 20 62.2 100 24.6 1,893
Oceania 7 92.5 100 15.5 470
Africa 4 90.2 91.0 17.6 91
Total 318 87.9 100 21.2 83,231

 

Note: 1) equity shares.

27

A Korean parent ﬁrm could ﬂexibly increase or decrease its share of affiliate’s
equity as Blodgett (1990) described, responding to the change in the affiliates’ speciﬁc
factors. In other words, the Korean multinational ﬁrm would actually change the equity
share of its affiliate as the relative importance of the intermediate inputs changed.
During 2002, some Korean parent ﬁrms in the data set invested a total of 2.868 billion
dollars to increase equity share of their overseas affiliates, and some Korean ﬁrms
withdrew a total of 184 million dollars through liquidating overseas affiliates or
decreasing equity ownership of their overseas affiliates. Thus, although we do not have
the panel data,29 we can test the hypothesis by using a cross section of Korean ﬁrm-
level data.

Among the 318 overseas affiliates, 166 affiliates belong to the manufacturing
industry and the other 152 affiliates are in non-manufacturing industry (Tables 1.1 and
1.9 in Appendix). In the manufacturing sector, the majority of overseas affiliates
(56.0%) are in the manufacturing industry of fabricated metal products, machinery and
equipment. In the non-manufacturing sector, 95 of 152 affiliates belong to the non-
manufacturing industry of wholesale and retail trade, and restaurants and hotels. Tables
1.2 and 1.10 in Appendix show that 66 and 57 overseas affiliates are located in China
and US, respectively. The majority of the Korean ﬁrrns’ overseas affiliates (167, 52.5%
of all overseas affiliates) are located in Asia.

In the data set, over 60% of all affiliates (197) are wholly owned (100 % ownership)
by their Korean parent ﬁrms (Table 1.3). Among these 197 afﬁliates, 87 affiliates

belong to the manufacturing industry. As a result, the ratio of wholly owned affiliates to

 

29 A parent ﬁrm might have never altered its equity share since it ﬁrst decided the share of affiliate’s
equity, even though the affiliate’s (or the parent ﬁrm’s) ratio of intangible assets to sales changed. Under
such circumstances, the panel data should be used to correctly test the relationship between parent ﬁrm’s
preference for equity ownership and the relative importance of intermediate inputs.

28

all afﬁliates in the non-manufacturing industry (72.3%) is much higher than that in the

manufacturing industry (52.4%).

Table 1.3: Year of Establishment and Afﬁliates by Equity Share

 

 

Estab. Years No. of affiliates Equity share No. of affiliates
Before 1990’s 74 s < 50% 19
1991-1996 148 50$s<75% 41
1997—2000 69 75 S s < 100% 61
2001-2002 27 s = 100% 197

 

 

To control the effects of host country’s income on ﬁrm’s equity ownership, we add
the host country’s per capita income. The host country’s per capita gross national
income of 2002 (per capita GNI) is obtained from the World Bank. We run regressions
by using the host country’s per capita gross national income in different ways to check
whether we obtain any inconsistent results. First, we run regressions by adding a control
variable for host country’s per capita gross national income. Second, instead of adding a
control variable, we divide the overseas afﬁliates into six groups with regards to host
country’s per capita gross national income level, and then use a dummy variable to
control the possibility that affiliates in higher income countries may have more
intangible assets (Tables 1.4 and 1.13 in Appendix).

In addition, to measure the institutional quality of each host country, the political
stability index developed by Kaufmann et al. (2002) is used. In the index, they combine
several indicators which measure perceptions of the likelihood that the government in
power will be destabilized or overthrown by possibly unconstitutional and / or violent
means, including terrorism. This index captures the idea that the quality of governance
in a country is compromised by the likelihood of wrenching changes in the government,

which not only have a direct effect on the continuity of policies, but also at a deeper

29

level, undermine the ability of all citizens to peacefully select and replace those in
power (Kaufmann et al., 2002, p. 5). This index ranges between -2.5 (lowest

institutional quality) and 2.5 (highest institutional quality).30 3'

Table 1.4: Afﬁliates by Income level" of Host Country

 

Income level No. of Affil. Average (%) Median (%) St. dev. Sales ($ mil.)

 

 

 

~ 700 25 83.3 100 21.1 2,295
700-1500 92 85.0 93.0 20.1 9,778
150010000 41 76.5 89.0 27.6 5,544
10000-20000 13 87.4 100 26.8 1,815
20000-27000 71 94.9 100 15.6 24,822
27000~ 69 92.8 100 19.1 38,976
Above 10000 153 93.2 100 18.4 66,236
Above 20000 140 93.7 100 ' 17.5 64,221

 

Note: 1) Per capita GNI (S). 2) Average, median and st. dev. are for equity shares.

To control the possibility that the differences in legal protections of investors affect
the Korean ﬁrm’s preference for equity participation, we categorize the host countries as
in La Porta et al. (1998). La Porta et al. classiﬁed 49 countries into four different legal
groups based on the origin of the country’s initial law, 32 such as English-origin
countries which are also called common-law countries, French-civil-law countries,

German-civil-law countries, and Scandinavian-civil-law countries.

 

3° For descriptive statistics, refer to Table 1.11 in Appendix.

3' The correlation matrix of the independent variables indicates little collinearity. Most correlations are
less than 0.37, and the highest correlation (between the host country’s per capita gross national income
and the host country’s political stability index) is 0.71.

32 The civil legal tradition originates in Roman law, uses statues and comprehensive codes as a primary
means of ordering legal material, and relies heavily on legal scholars to ascertain and formulate its rules.
Legal scholars typically identify three currently common families of laws within the civil-law tradition:
French, German, and Scandinavian. However, the common law is formed by judges who have to resolve
speciﬁc disputes. Precedents from judicial decisions shape common law (La Porta et al., 1998, pp. 1118-
1119).

30

In La Porta et al.’s paper, relatively higher ownership concentration is found in the
countries where investors were poorly protected. They showed that good accounting
standards and shareholder protection measures are associated with a lower ownership
concentration. Based on their classiﬁcations of countries, we add dummy variables to

represent the differences in legal protections of investors (Table 1.5).

Table 1.5: Countries by Legal Origin

 

Origins Countries

 

Common-law Australia, Canada, Hong Kong, India, Ireland, Malaysia, Singapore,

South Africa, Sri Lanka, Thailand, U.K., U.S., Israel.

 

French-civil-law Argentina, Belgium, Brazil, Chile, France, Indonesia, Italy, Mexico,

Netherlands, Peru, Philippines, Portugal, Spain, Turkey.

 

German-civil-
Austria, Germany, Japan, Switzerland, Taiwan.

 

 

 

Law
Scandinavian-
Sweden.
civil-law
Others Bangladesh, China, Vietnam, Uzbekistan, Czech Rep., Hungary,
Poland, Romania, Russia, Cayman Islands, Panama, Puerto Rico,
Virgin Islands, Guam, Morocco, Sudan.
3.3 Results

We divide the whole sample into two sub-samples such as a sample of
manufacturing industry, and a sample of non-manufacturing industry. The dependent
variable is the logarithm value of equity share (quuiry) , and the independent variables

are the ratio of intangible assets to sales of parent ﬁrm (lpit), the ratio of intangible

31

assets to sales of affiliate (lsit) , the logarithm value of parent ﬁrm’s sales (lpsale), the
logarithm value of affiliate’s sale (lssale), host country’s political stability (stabil),
host country’s per capita gross national income (lpgm'), and dummies for affiliate’s

location and industry (dummies).

Table 1.6: Equity Ownership Structures with Location Dummy

 

 

Whole industry Manufacturing Non-manufacturing
[pit -0.000 -0.002 -0.002
(0.01 1) (0.014) (0.020)
Isit -0.048*** -0.045*** -0.048*
(0.01 1) (0.012) (0.024)
lpsale -0.020 -0009 -0.010
(0.013) (0.016) (0.027)
lssale -0.03 l * -0.063*** -0.036
(0.017) (0.023) (0.025)
stabil 0.059 0.068 -1.003***
(0.046) (0.044) (0.335)
lpgni 0051* 0.042 0340*“
(0.029) (0.039) (0.088)
dummies for location and industry.
No. of obs. 134 88 46
R2 0.39 0.34 0.67

 

Note: Standard errors in parenthesis.
(*, and *" are 10%, and 1% signiﬁcance levels, respectively.)

The empirical results partially support the theory. Table 1.6 shows that the lsit is

statistically signiﬁcant in the whole industry and manufacturing sector at the 1% level,

32

and it is also signiﬁcant at the 10% level in the non-manufacturing sector. However, the
[pit is not statistically different from zero. That is, if an affiliate has a higher ratio of
intangible assets to sales, then the parent firm of the affiliate gives more of the equity
share to its affiliate. However, the empirical results show that the parent ﬁrm’s ratio of
intangible assets to sales does not play any important role in parent ﬁrm’s choice of
equity participation.33

Another interesting result is that the effect of an affiliate’s size (lssale) on the parent
ﬁrm’s equity ownership is signiﬁcantly negative in the manufacturing sector. A Korean
parent ﬁrm holds less of the equity share when its overseas afﬁliate is very big. It is
because the parent ﬁrm might want to reduce the risks to which it is exposed, through
choosing a joint venture. The other possible reason is that the host country imposes
some restrictions on foreigner’s equity ownership when the local ﬁrm is very big. The
host country might be concerned about the possibility of future unemployment when the
foreign parent ﬁrm withdraws its investment. However, the size of the parent ﬁrm
(lpsale) does not play any important role in the parent ﬁrm’s equity ownership.

The sign of stabil is negative at the 1% signiﬁcance level in the non-
manufacturing sector. Later, we discuss the effects of host country’s political stability on
the ﬁrm’s preference for equity participation. The coefficient on lpgni is positive and
statistically signiﬁcant in the whole industry (10% level) and in the non-manufacturing
sector (1% level). This shows that Korean ﬁrms in the non-manufacturing sector tend to
hold greater equity shares of their affiliates when they invest in higher income level

countries. The qualitative results do not change when we alter the categorization of the

 

’3 We also run regressions by using the logarithm value of parent ﬁrm’s intangible assets ( plin) and the
logarithm value of affiliate’s intangible assets (slin) instead of [pit and [sit . Similarly, however,
plin is not statistically signiﬁcant, but slin is signiﬁcantly negative. In addition, we use the logarithm

value of the ratio of affiliate’s intangible assets to parent ﬁrm’s intangible assets (Iratio ). The sign of
[ratio is signiﬁcantly negative both in the whole industry and in the manufacturing industry, but is not
signiﬁcant in the non-manufacturing industry.

33

host countries. That is, we run regressions again affer we put the host countries in
Oceania34 into Asia and / or divide Europe into Western Europe and Eastern Europe.
The sign of Isit is still signiﬁcantly negative irrespective of the changes in

categorization of the host countries.35

Table 1.7 : Equity Ownership Structures (above $10,000 income)

 

Whole industry Manufacturing Non-manufacturing
lpit -0.008 -0034 -0002
(0.016) (0.026) (0.022)
Isit -0.053*** -0.082** -0.049**
(0.015) (0.021) (0.021)
lpsale 0044" -0.048 -0039
(0.022) (0.040) (0.028)
lssale -0.041 * -0.074 -0.041
(0.021) (0.054) (0.030)
stabil -l.045*** -0.546 -1.139*"‘*
(0.256) (0.604) (0.348)
lpgni 0353*" 0.510" 0382*
(0.125) (0.185) (0.368)
No. ofobs. 54 17 37
R2 0.45 0.80 0.37

 

Note: Standard errors in parenthesis.
(*, *"', and ”'" are 10%, 5%, and 1% signiﬁcance levels, respectively.)

When we add the dummy for host country’s legal orientation to control the

 

3" Australia and Guam.
3’ When we run regressions with partially owned affiliates (pure joint ventures) by excluding 100%
owned affiliates, we cannot ﬁnd any meaningful results.

34

possibility that the differences in investor protections might affect the multinational
ﬁrm’s equity ownership structure, the Isit is still negative and statistically signiﬁcant
(Table 1.12 in Appendix). In addition, we run regressions again by categorizing
afﬁliates into six groups with respect to the host country’s income level as in Table 1.4,
to check whether we can obtain different results. In each group, host countries whose
per capita gross national incomes are correspondent to the given range are included.
Including the dummy variables for the income level and industry, we obtain the
analogous results that the Isit is statistically signiﬁcant (Table 1.14 in Appendix). In
contrast to Table 1.6, Table 1.14 shows that the lssale is negative and statistically
signiﬁcant at the 10% level only in the manufacturing sector.

Considering the possibility that the affiliates located in higher income countries
have meaningful intangible assets, we test the same hypothesis with the affiliates
located in countries where the per capita gross national incomes are above 10,000 US.
dollars (Table 1.7). In this small sample, the Isit is still statistically signiﬁcant and
negative. With the affiliates which are located in much higher income level (above
20,000 US. dollars), the lsit is also statistically signiﬁcant and negative (Table 1.8).
However, the [pit does not change to be signiﬁcant.36

It is expected that if the host country is politically more stable, then a parent ﬁrm
holds a greater share of affiliate’s equity. A parent ﬁrm might heavily depend on its local
partner since the local partner can play important roles in handling the problems
originating from the host country’s instability. Hence, the Korean ﬁrm might tend to
hold a smaller equity share when it acquires an affiliate in a more unstable country. This
expectation is supported since the stabil is signiﬁcantly positive in the manufacturing

sector in Tables 1.12 and 1.14 in Appendix. However, Table 1.6 shows that the stabil

 

3" The independent variables, [pit and [sit might be endogenous. Thus, the causality might go ﬁ'om

ownership to the ratios of intangible assets to sales. However, we cannot reject that the causality goes
from ratios of intangible assets to sales to ownership since we do not have instrument variables.

35

is statistically negative in the non-manufacturing sector. In addition, Tables 1.7 and 1.8
show that the signs of stabil are signiﬁcantly negative in the whole and non-
manufacturing industry. Thus, we cannot ﬁnd any consistent results concerning the

effects of host country’s political stability on ﬁrm’s equity ownership.

Table 1.8: Equity Ownership Structures (above $20,000 income)

 

 

Whole industry Manufacturing Non-manufacturing
1121'! -0.000 -0.026 0.004
(0.016) (0.024) (0.024)
lsit -0.066*** -0.071** 0072*"
(0.015) (0.022) (0.023)
lpsale -0.061* -0037 -0.066"‘*
(0.023) (0.038) ' (0.032)
lssale -0022 -0.013 -0.026
(0.021) (0.063) (0.030)
stabil -1.733*** -1.421 -1.886***
(0.347) (0.870) (0.487)
lpgni 0591*" 0.616" 0.574M
(0.172) (0.220) (0.263)
No. of obs. 46 15 31
R2 0.59 0.86 0.53

 

Note: Standard errors in parenthesis.

(*, ", and ”* are 10%, 5%, and 1% significance levels, respectively.)

To capture the possibility that Korean economic difficulties might affect the Korean
firms’ equity ownership structures, we add year dummy variables for the periods of

Korean economic problems. There have been two different years of major economic

36

turmoil in Korea since a Korean parent ﬁrm ﬁrst invested in an overseas affiliate in
1973. In 1980, Korea faced many difficulties throughout the entire economy. These
included its ﬁrst ever negative grth since planned development was launched in 1962.
Also Korea experienced a foreign exchange crisis in 1997. To represent these economic
difficulties, two year dummies for 1980 and 1997 are used to control these uncommon
years. However, those dummies are not statistically signiﬁcant and the qualitative

results do not change. 37

4. Conclusion and Further Studies

We examine the multinational ﬁrm’s vertical integration and show that joint
ventures are the multinational ﬁrm’s organizational choices under the situation of ex
ante incomplete contracts. In order to make a ﬁnal product, it is assumed that a
multinational ﬁrm in the source country produces an intermediate input and purchases
another intermediate input from a supplier in the host country to assemble the two
inputs. However, the quality of intermediate input is assumed to be non-contractible ex
ante. To produce high-quality intermediate input, the multinational ﬁrm and the supplier
are assumed to use speciﬁc factors.

We ﬁnd that if the supplier’s intermediate input is important, then the multinational
ﬁrm (ﬁnal-product producer) holds a smaller equity share of supplier’s ﬁrm to give the
supplier an incentive to produce high-quality intermediate input. However, if the
intermediate input made by the producer is more important than the input made by

supplier, then the producer is inclined to hold a greater equity share of supplier’s ﬁrm to

 

’7 As in the following section, we also run regressions aﬁer adding other control variables to represent the
characteristics of the host countries, such as the measurement of intellectual property rights, the index for
human capital abundance, the host country’s purchasing-power parity adjusted gross domestic product,
telephone mainlines per 1000 inhabitants, the host country’s openness to foreign direct investment, and an
indicator of the share of black market economy. None of these aforementioned control variables are
statistically signiﬁcant, which might be caused by a lack of observations in the data set used in this
section. However, the sign of Isit is still statistically signiﬁcant although we add all these variables.

37

 

take a larger fraction of total revenues. We also show that intermediate inputs are always
produced at a level lower than the socially optimal production levels. In addition, we
theoretically ﬁnd that the producer decreases its share of supplier ﬁrm’s equity as the
market becomes more competitive.

Using Korean ﬁrm-level data, we test the prediction in the theory that the
multinational ﬁrm chooses its equity share of supplier’s ﬁrm depending on the relative
importance of intermediate inputs. To represent speciﬁc factors in the theory, we use the
ﬁrrn’s intangible assets; such as a brand, franchise, or patent. The empirical tests
partially support the theory that if an affiliate has the higher ratio of intangible assets to
sales, then its Korean multinational ﬁrm is inclined to hold a smaller share of the
affiliate’s equity. However, the multinational ﬁrm’s ratio of intangible assets to sales is
not statistically signiﬁcant.

That means there might be other factors which affect the multinational ﬁrm’s choice
of an equity share. There are many things leff to be done to explain the boundaries of
the ﬁrm. We do not consider the possibilities that the two parties, the producer
(multinational ﬁrm) and its intermediate-input supplier (affiliate), come into conﬂicts of
interests. These conﬂicts are strongly related to issues concerning multinational ﬁrm’s
management costs.

In addition, technology leakage from multinational ﬁrms to affiliates or to other
local companies in the host country is also important. Technology leakage is a very
sensitive problem to a multinational ﬁrm since a multinational firm is usually an
oligopolist in its industry and has advanced technology (Markusen, 1995). Due to the
technology leakage problem, a multinational ﬁrm might prefer whole ownership to
outsourcing or a joint venture. In another chapter, we consider an important practical
question for multinational ﬁrm, what entry mode a multinational ﬁrm chooses if the
technology leakage risk is not independent of its entry mode?

38

 

APPENDIX

Table 1.9: List of Industries

 

Industries

 

Manufacturing Food, Beverages and Tobacco (9), Textile, Wearing Apparel

(166)

and Leather (26), Wood and Wood Products Including
Furniture (2), Paper and Paper Products, Printing and
Publishing (2), Chemicals and Chemical, Petroleum, Coal,
Rubber and Plastic Products (16), Non-Metallic Mineral
Products except Products of Petroleum and Coal (3), Basic
Metal Industy (l3), Fabricated Metal Products, Machinery

and Equipment (93), Other Manufacturing Industries (2).

 

Non-
Manufacturing

(152)

Mining and Quarrying (10), Electricity, Gas and Water (2),
Construction (8), Wholesale & Retail Trade and Restaurants
& Hotels (95), Transport, Storage and Communication (6),
Financing, Insurance, Real Estate and Business Services (24),

Community, Social and Personal Services (7).

 

Note: Numbers of overseas afﬁliates in parenthesis.

39

Table 1.10: Countries by Location

 

 

 

 

 

 

 

Location Countries
Japan (11), Bangladesh (1 ), Sri Lanka (1), India (5),
Hong Kong (19), Indonesia (1 7), Israel (1), China (66),
Asia (167)
Malaysia (4), Taiwan (1), Philippines (7), Singapore (9),
Thailand (7), Vietnam (15), Uzbekistan (3).
North America (61) Canada (4), US. (5 7).
Belgium (2), France (4), Germany (14), Ireland (1),
Italy (3 ), Netherlands (5), Turkey (2), Portugal (1),
Europe (59) Spain (2), UK. (11), Austria (1), Sweden (1),
Switzerland (1), Czech Rep. (1), Hungary (4),
Poland (2), Romania (2), Russia (2).
Bermuda (1), Argentina (1), Brazil (4), Chile (1 ),
South America (20) Mexico (5), Peru (1), Cayman Islands (1 ), Panama (3),
Puerto Rico (1), Virgin Islands (2).
Oceania (7) Australia (5), Guam (2).
Africa (4) South Africa (2), Morocco (1), Sudan (1).

 

Note: Numbers of overseas affiliates in parenthesis.

4O

Table 1.11: Descriptive Statistics

 

 

 

 

Obs. Mean St. dev. Min. Max

equity 318 87.9 21.2 5.00 100

lpit 293 -490 2.01 -1325 -0.62

lsit 156 -474 2.57 -10.93 2.57

Whole
lpsale 314 14.76 1.96 8.52 17.51
Industry

lssale 283 10.91 2.12 1.38 16.10

stabil 312 0.60 0.76 -2.01 1.61
lpgni 311 8.58 1.69 5.73 10.54

equity 166 88.0 17.8 10.0 100

lpit 153 -4.64 1.99 -13.24 -0.61

[sit 101 -4.55 2.30 -9.69 2.57

Manufac. lpsale 166 14.40 1.88 8.52 17.51
lssale 153 11.06 1.58 3.91 15.29

stabil 164 0.35 0.78 -2.01 1.48
lpgni 163 7.86 1.57 5.73 10.46

equity 152 87.8 24.5 5.00 100

lpit 140 -5.17 2.00 -10.64 0.68

[sit 55 -5.10 2.99 -1092 0.68

Non-
lpsale 148 15.17 1.97 8.97 17.51
Manufac.

lssale 130 10.72 2.60 1.38 16.10

stabil 148 0.88 0.62 -1 .56 1.61

lpgni 148 9.38 1.44 5.88 10.54

 

41

Table 1.12: Equity Ownership Structures with Legal Origin Dummy

 

 

Whole industry Manufacturing Non-manufacturing
1172'! 0.002 0.006 -0003
(0.011) (0.014) (0.019)
Isit -0.039*** -0.027* * -0.052**
(0.011) (0.012) (0.024)
lpsaie -0.018 -0017 -0021
(0.014) (0.017) (0.026)
lssale -0.028 -0.044* -0.026
(0.01 8) (0.024) (0.024)
stabil 0.091 0.160“ -1 397*“
(0.063) (0.061) (0.339)
lpgni -0.028 -0.081* 0.274"
(0.042) (0.042) (0.109)
dummies for legal origin and industry.
No. of obs. 134 88 46
R2 0.32 0.23 0.69

 

Note: Standard errors in parenthesis.

(‘, ", and *** are 10%, 5%, and 1% signiﬁcance levels, respectively.)

42

Table 1.13: Countries by Income Level”

 

Income Level

Countries

 

Vietnam (15), Sudan (1), India (5), Uzbekistan (3),

 

 

 

 

 

~ 700 (25)
Bangladesh (1).
Indonesia (1 7), Philippines (7), China (66), Sri Lanka (1),
700~1500 (92)
Morocco (1).
Panama (3), Thailand (7), Malaysia (4), Hungary (4),
Mexico (5), Russia (2), Chile (1), Poland (2), Brazil (4),
1500~10000(41)
South Africa (2), Romania (2), Argentina (1), Turkey (2),
Czech Rep. (1), Peru (1).
Australia (5), Italy (3), Israel (1), Spain (2), Portugal (1),
10000~20000 (13)
Puerto Rico (1).
Germany (14), UK. (11), Hong Kong (19), Singapore (9),
20000~27000 (71) Canada (4), Belgium (2), France (4), Sweden (1),
Ireland (1), Netherlands (5), Austria (1).
27000~ (69) US. (57), Japan (1 1), Switzerland (1).

 

Note: 1) Per capita GNI (8). Numbers of overseas affiliates in parenthesis.
2) Per capita Gle for Guam (2), Taiwan (1), Bermuda (1), Cayman Islands (1),

Virgin Islands (2) are not available.

43

Table 1.14:

Equity Ownership Structures with Income Level Dummy

 

Whole industry Manufacturing Non-manufacturing
1121'! -0.002 0.009 -0.0117
(0.011) (0.014) (0.020)
[sit -0.039*** -0.032** -0.051**
(0.010) (0.013) (0.021)
lpsale -0.020 -0017 -0015
(0.014) (0.018) (0.028)
lssale -0.028 -0.045* -0.030
(0.018) (0.025) (0.026)
stabil 0.063 0.090" -0.049*
(0.044) (0.044) (0.025)
dummies for income level and industry.
No. of obs. 136 89 47
R2 0.33 0.21 0.66

 

Note: Standard errors in parenthesis.

C“, ", and "* are 10%, 5%, and 1% signiﬁcance levels, respectively.)

44

Figure 1.2: Production Process

 

 

     

Final-Product Producer

       
     
   
 
 
   

produces xS

    

Final-Product Producer

produces final product y

     

by using x8 and x”

Intermediate-Input Supplier

 

produces x”

 

 

 

 

 

 

 

 

 

.15lllll'l'llllllllllllllllllllllll
1.1.1.
.051
0 2....-___- 1' ”N'ﬂ-‘M‘
° 1 21! Hi 1401144 siil

 

 

 

 

 

 

 

 

 

 

Figure 1.3: Density of Overseas Affiliates

45

REFERENCES

Aghion, Philippe, and Jean Tirole (1994) “The Management of Innovation,” Quarterly
Journal of Economics, 109(4), 1185-1209.

Antras, Pol (2003) “Firms, Contracts, and Trade Structure,” Quarterly Journal of
Economics, 118(4),1374-1418.

, and Elhanan Helpman (2004) “Global Sourcing,” Journal of Political
Economy, 112, 553-580.

Behrman, Jack N., and Robert E. Grosse (1990) International Business and
Governments: Issues and Institutions, University of South Carolina Press.

Blodgett, Linda Longfellow (1991) “Partner Contributions as Predictors of Equity Share
in International Joint Ventures,” Journal of International Business Studies, 22 (I),
63-78.

Blomstrdm, Magnus, and Ari Kokko (1997) “The Impact of Foreign Investment on Host
Countries: A Review of the Empirical Evidence,” World Bank Policy Research
Working Paper #1745.

Caves, Richard E. (1996) Multinational Firm and Economic Analysis, Cambridge:
Cambridge University Press.

Crampton, Peter, Robert Gibbons, and Paul Klemperer (1987) “Dissolving a Partnership
Efficiently,” Econometrica, 55 (3), 615-632.

Desai, Mihir A., C. Fritz Foley, and James R. Hines Jr. (2004) “The Costs of Shared
Ownership: Evidence from International Joint Ventures,” Journal of Financial

Economics, 73 (2), 323-374.

The Export-Import Bank of Korea (2003) Financial Statement Analysis of Koreas
Foreign Subsidiaries, The Export-Import Bank of Korea.

Fagre, Nathan, and Louis T. Wells Jr. (1982) “Bargaining Power of Multinationals and
Host Governments,” Journal of International Business Studies, 13 (2), 9-23.

46

Gatignon, Hubert, and Erin Anderson (1988) “The Multinational Corporation’s Degree
of Control over Foreign Subsidiaries: An Empirical Test of a Transaction Cost
Explanation,” Journal of Law, Economics, and Organization, 4 (2), 304-336.

Grossman, Sanford J., and Oliver D. Hart (1986) “The Costs and Beneﬁts of
Ownership: A Theory of Vertical and Lateral Integration,” Journal of Political
Economy, 94 (4), 691-719.

Grossman, Gene M., and Elhanan Helpman (2002) “Integration versus Outsourcing in
Industry Equilibrium,” Quarterly Journal of Economics, 117 (1), 85-120.

Hart, Oliver (1995) Firms, Contracts, and Financial Structure, Oxford: Clarendon Press.

Hennart, Jean-Francois (1991) “The Transaction Costs Theory of Joint Venture: An

Empirical Study of Japanese Subsidiaries in the United States,” Management
Science, 37 (4), 483-497.

Kaufrnann, Daniel, Aart Kraay, and Pablo Zoido-Lobaton (2002) “Governance Matters
2: Updated Indicators for 2000/01,” World Bank Policy Research Working Paper
#2772.

Kimura, Yui, and Thomas A. Pugel (1995) “Keiretsu and Japanese Direct Investment in
US Manufacturing,” Japan and the World Economy, 7, 481-503.

Klein, Benjamin, Robert G Crawford, and Armen A. Alchian (1978) “Vertical
Integration, Appropriable Rents, and the Competitive Contracting Process,”
Journal of Law and Economics, 21, 297-326.

Kobrin, Stephen J. (1987) “Testing the Bargaining Hypothesis in the Manufacturing
Sector in Developing Countries,” International Organization, 41 (4), 609-63 8.

La Porta, Rafael, F lorencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny
(1998) “Law and Finance,” The Journal of Political Economy, 106(6), 1113-1155.

Lecraw, Donald J. (1984) “Bargaining Power, Ownership, and Proﬁtability of
Multinational Corporations in Developing Countries,” Journal of International
Business Studies, 15 (1), 27-43.

Mataloni, Raymond J. Jr. (2002) “US. Multinational Companies: Operations in 2000,”

47

Survey of Current Business, 82, 1 11-131.

Markusen, James R. (1995) “The Boundaries of Multinational Enterprises and the
Theory of International Trade,” Journal of Economic Perspectives, 9 (2), 169-189.

Svejnar, Jan, and Stephen C. Smith (1984) “The Economics of Joint Ventures in Less
Developed Countries,” Quarterly Journal of Economics, 99 (1), 149-167.

Tirole, Jean (1988) “The Theory of the Firm,” The Theory of Industrial Organization,
Cambridge MA: MIT Press.

Whinston, Michael D. (2003) “On the Transaction Cost Determinants of Vertical
Integration,” Journal of Law, Economic, and Organization, 19(1), 1-23.

Wooldridge, Jeffrey M. (2000) Introductory Econometrics - A Modern Approach, South-
Westem College Publishing.

48

CHAPTER II

Firm’s Ownership Structures: Intellectual Property
Protection, Resources-Based Sector,

and Market Competition

1. Introduction

How international differences in intellectual property rights protection affects the
multinational ﬁrm’s decision concerning ownership structure is an important question.
Multinational ﬁrms have important intangible assets (or superior knowledge) and they
worry about leakage of these assets when entering foreign countries. The multinational
ﬁrms that invested heavily in creating new knowledge have strong incentives to control
the diffusion of their hard-earned assets. They can take action to limit leakage of their
assets. For example, they consider the expected costs of intangible assets’ diffusion
while making decisions regarding where to set up affiliates.

Faced with this diffusion, a multinational ﬁrm often chooses to exploit its ﬁrrn-
speciﬁc advantages by internalizing the transactions through wholly-owned affiliates
rather than via joint ownership or licensing. We analyze with ﬁrm-level data the
relationship between the host country’s standards of intellectual property rights and
multinational ﬁrms’ equity participations in its foreign affiliates. That is, we try to
answer an important question: whether a multinational ﬁrm is inclined to have a greater
equity share when host country’s intellectual property protection is weak. 1

After transferring its superior knowledge to a foreign affiliate, a multinational ﬁrm

has much more difficulty preventing local companies in the host country from
49

appropriating the superior knowledge when the ﬁrm owns the affiliate partially than
when the ﬁrm owns the affiliate wholly. For example, multinational ﬁrms fear that if
local partners control the employment policy, they may not put enough effort into
keeping key employees, who may leave and reveal their, knowledge of production
processes (Smarzynska, 2000). Thus, in order to guard against leakage of important
technologies, multinational ﬁrms prefer wholly-owned affiliates to shared ownership.1

In the following section, we review related literature. However, most existing
empirical studies focus on technology (intangible assets) spillovers into the host
countries, rather than the multinational ﬁrm’s reactive level of equity participation in its
affiliate relative to the degree of intellectual property protection in the host country. In
addition, most papers on multinational ﬁrms consider the effects of intellectual property
rights on the choice between whole-ownership of affiliates and licenses, or between
whole-ownership of affiliates and exports. As a result, there is little evidence that a
multinational ﬁrm’s equity participation in its affiliate is negatively related to the host
country’s standards of intellectual property rights protection.

We try to answer the essential question on the ﬁrm’s boundaries: whether a
multinational ﬁrm really changes its equity structure according to the host country’s
standards of intellectual property rights protection. If the ﬁrm does change it, which
direction it is inclined to go? We examine this issue by using Korean ﬁrm-level data. We
ﬁnd that weak protection of intellectual property rights induces a multinational ﬁrm to
increase its equity share of affiliate. Furthermore, the ﬁrrn-Ievel data set with a large
number of observations allows us to examine more deeply the relationship between the

degree of intellectual property rights protection and the multinational ﬁrm’s preferences

for equity participation.

 

' Shared ownership (or joint ownership) means that residual control rights of an affiliate are shared by a
multinational ﬁrm and the ﬁrm’s local partner.
50

Although the Korean ﬁrm-level data set includes ﬁrms that established affiliates as
early as 1973, most investment has occurred since 1990 when the average technologies
of Korean ﬁrms began to be comparable with those of other developed countries2 (Table
2.3). In addition, the majority of Korean ﬁrms’ overseas affiliates are located in Asia,
where the economies of most of the countries are less developed3 than the Korean
economy (Tables 2.1 and 2.18 in Appendix).

In 2003, the Ministry of Commerce, Industry, and Energy of Korea surveyed more
than 650 Korean manufacturing ﬁrms which invested'in affiliates located in Asia.4 The
survey results show that more than thirty percent of ﬁrms transferred their core
technologies to the overseas affiliates, and more than forty ﬁve percent of Korean ﬁrms
transferred their non-core technologies to overseas affiliates. Hence, Korean
multinational ﬁrms would be concerned about the leakage of their technologies when
they invested in foreign countries.

Based on the survey data of Indonesian establishments in 1991, Blomstrom and
Sjoholm (1999) raised an interesting question: “Does the degree of spillover differ with
the degree of ownership in foreign direct investment?” However, their main focuses are
technologies spillovers and ﬁrms’ productivity.5 In contrast to their paper, we test the
relationship between a Korean ﬁrm’s equity participation in its foreign affiliates and the
standards of intellectual property rights protection in the host country. The ﬁrm-level

data set is broad, and it is disaggregated into industry levels (19 manufacturing

 

2 It is not unreasonable to assume that the Korean ﬁrms which invested in overseas affiliates before the
19905 had important intangible assets in some areas compared to local competitors in host countries. In
addition, Korean ﬁrms which invested in developed countries, such as U.S., U.K., and Japan, might have
more advanced technologies in some speciﬁc areas than other local competitors in those developed
countries. Thus, we can speculate that Korean ﬁrms would be concerned about the leakage of their
advanced technologies to the local competitors when they decided to invest in foreign countries.
3 Among 7,635 affiliates in Asia, only 113 affiliates are located in Japan (Table 2.1).
4 The Ministry of Commerce, Industry, and Energy of Korea (2003), “The Survey Results on Korean
Manufacturing Firms’ Overseas Direct Investments.”
5 BlomstrOm and Sj6holm ( 1999) found that the degree of foreign ownership affected neither the level of
labor productivity nor the extent of spillovers in Indonesian manufacturing industries. Using the data set
of ﬁrms in Greece, however, Dimelis and Louri (2002) showed that the higher the degree of foreign
ownership, the more efficient is production in Greece.

51

industries in 98 countries). This allows us to analyze the relationship in various ways,
and make other interesting hypotheses related to ﬁrm ’3 preference for equity ownership.
We examine the effects of resources-base sector on the multinational ﬁrm’s equity
participation in its affiliate. We test whether a multinational ﬁrm increases its equity
share when its affiliate is in a host country where the legal protection of shareholders is
weak. In addition, we study whether a multinational ﬁrm prefers joint ownership when
it invests in the resources-based sector. Although many past studies by Gatignon &
Anderson (1988), and Gomes-Casseres (1989, 1990), have examined the effects of
resources-based sector, they used only data on US. multinational ﬁrms. However, we
analyze the choices of multinational ﬁrms in a newly developed country (Korea).
Finally, to show whether there exists any causal connections between foreign direct
investment and market concentration, we test the relationships6 between the market
competition of an industry and the equity participation in an affiliate of the industry.
This study is structured as follows. The following section brieﬂy reviews the related
literature and formulates the hypotheses to be tested. Section 3 describes the data set
and statistical model. In section 4, empirical results are presented and we conclude the

study in section 5.

2. Related Literature and Hypotheses

The linkage between the technology spillovers to the host country and foreign direct
investment has been deeply analyzed. Blomstrom and Kokko (1997) surveyed the
empirical evidence on the impact of foreign direct investment on host countries. Their
main conclusion was that a multinational ﬁrm may play an important role for

productivity and export growth in the host country. However, papers on the relationship

 

6 This relationship was analyzed theoretically in the previous chapter; “Outsourcing, Joint Ventures, and
Whole Ownership.”
52

between the host country’s intellectual property protection and multinational ﬁrm’s
equity participation to foreign affiliates are relatively rare.

Theoretically, Nicholson (2002) showed that as intellectual property rights in the
South strengthen, firms in the North are more likely to shift production to the South. In
addition, ﬁrms that face a lower risk of imitation, or are less technically advanced, will
tend to license production to non-affiliated Southern ﬁrms. Vishwasrao (1994)
demonstrated that the lack of adequate intellectual property protection may encourage
foreign direct investment relative to licensing. The trade-off between foreign direct
investment and licensing is that foreign direct investment avoids the risk of imitation at
the expense of higher costs.

Markusen (2001) demonstrated that multinational ﬁrms choose to acquire affiliates
in equilibrium when the intellectual property protection is low, emphasizing contract
enforcement concerns. Ethier and Markusen (1996) explained that in the absence of
intellectual property rights protection, ﬁrms internalize their technology transactions to
avoid the dissipation of their property assets. They showed that if knowledge capital is
of medium or high importance relative to physical capital, then ﬁrms in the source
country would choose to be multinational firms. In addition, Markusen (1995) explained
that low standards of intellectual property rights may be associated with foreign direct
investment being chosen as the governance structure, although the costs of doing
business abroad is higher than the costs of licensing.

To protect the ﬁrm’s technology, whether or not wholly-owned affiliates are
preferable to joint ownership, is an important issue. Several host countries were known
to require multinational ﬁrms to take on local partners to encourage technology
spillovers. Joint ownership might be avoided by the multinational ﬁrms that have
appropriable proprietary assets. By using selected samples of US. based ﬁrms,

Mansﬁeld and Romeo (1980) showed that the incentive to prevent the dissipation of
53

 

advanced technologies is reﬂected in the fact that multinational ﬁrms transfer their
technologies of new vintage via direct investment, preferring this to licensing or
forming joint ownership, but transfer their old technologies through joint ownership.
However, they used only thirty one U.S.-based ﬁrms to test their hypotheses.

Gomes-Casseres (1989) illustrated that multinational ﬁrms contribute less complete
technological packages to joint ventures than they do to their wholly-owned affiliates.
By using US. ﬁrm-level panel data, Branstetter et al. (2005) showed that multinational
ﬁrms respond to stronger intellectual property rights regimes by increasing their
technology transfer to the reforming countries. Their ﬁndings are consistent with either
an increase in the volume of technology being transferred or an increase in the degree to
which US. multinational ﬁrms can extract higher level of rent from technologies that
have already been deployed in the host countries.

Mansﬁeld (1994 and 1995), and Lee and Mansﬁeld (1996), found that the host
country’s intellectual property protection offen has a signiﬁcant effect on the amount
and kinds of technology transfer and direct investment, based on survey data, interview
studies, and statistical analysis. According to Smarzynska’s study by using ﬁrm-level
data of Eastern Europe and the former Soviet Union (2000), multinational ﬁrms that are
technological leaders in their industries are more likely to wholly own their affiliates
than to share ownership with their local partners. In her later paper (2004), she showed
that foreign direct investors in sectors relying heavily on protection of intellectual
property are discouraged by weak intellectual property rights. However, her data set
contains information only on Eastern Europe and the former Soviet Union. None of
these studies have examined the effects of intellectual property rights regime on the
composition of equity structure.

In order to test the relationship between legal protection of investors and ownership

concentration, La Porta et al. (1998) examined legal rules covering protection of ﬁrms’
54

shareholders in 49 countries and assembled a database of the 10 largest non-ﬁnancial
domestic ﬁrms in each country. They found that good accounting standards and
shareholder protection measures are associated with a lower concentration of ownership.
In contrast to La Porta et al.’s paper that focused on domestic firms’ ownership
structures, we study multinational ﬁrms’ ownership preferences and test whether a
multinational ﬁrm increases its equity share when its affiliate is located in a country
where the legal protection of shareholders is weak.

Recently, McCalman (2004) analyzed the relationship between intellectual property
rights and foreign direct investment. He used the governance structure of the operations
of major Hollywood studios in 40 foreign markets in 1997. He found that both high and
low standards of intellectual property rights encourages more integrated governance
structures, although moderate intellectual property rights are assOciated with a high
degree of licensing. However, there are two important differences between our study
and his. First, our analysis is based on production processes, different from his basis of
distribution processes. Second, he considered only two extreme cases: foreign direct
investment and licensing. We examine whether a multinational ﬁrm raises its equity
share as the host country’s intellectual property rights protection becomes weak.

We concentrate on the multinational ﬁrm’s reaction to host country’s standards of
intellectual property rights protection. The main hypothesis is that the weaker the
standards of intellectual property rights protection in the host country, the higher an
equity share a multinational ﬁrm holds. Although some multinational ﬁrms would
decide their entry modes in one stage, other multinational ﬁrms would decide their
choices of entry mode in two stages. In the ﬁrst stage, the multinational ﬁrms decide
between whole ownership and a joint venture. In the second stage, the multinational
ﬁrms choose equity participation in their affiliates if they decided on a joint venture

instead of whole ownership as an entry mode in the ﬁrst stage. To encompass these
55

possibilities, we use two different samples, a sample of all overseas afﬁliates, and a
sample of pure joint ventures, to test the hypothesis.

We also test the relationship between the host country’s investor protection and a
multinational ﬁrm’s equity participation. In addition, we test other issues on the host
countries’ characteristics. We try to answer whether a multinational ﬁrm prefers joint
ownership to complete ownership when its affiliate is in resources-based industries. We
have another theoretical prediction in the previous chapter that when the market
becomes more competitive, then a multinational ﬁrm decreases its equity participation
in its affiliate. We can test the prediction due to the larger amount of observations in the

data set used in this chapter, compared to the data set used in the previous chapter.

3. Data and Statistical Model

We take the ﬁrm-level data used from the Overseas Direct Investment Statistics
Yearbook published by the Export-Import Bank of Korea. With this data set, we classify
every overseas affiliate into a 2-digit Korean Standard Industrial Classification code.7
However, we have only limited information on the Korean frrms’ overseas direct
investments; such as overseas (host) countries, Korean investors, foreign affiliates,
equity shares which Korean investors hold, 2-digit Korean Standard Industrial
Classification codes for each affiliate, investment amounts, and the year of investment.
We have 19 sub-manufacturing industries which are classiﬁed by the Korean Standard
Industrial Classiﬁcation code (Table 2.16 in Appendix).

Table 2.1 states that the majority of Korean ﬁrms’ overseas affiliates are located in
Asia. China is the most attractive host country for Korean multinational firms. In

addition, most Korean multinational ﬁrms’ overseas joint ventures are also located in

 

7 We gratefully thank Junghwa Seo and James Kim of the Export-Import Bank of Korea for helping to
reﬁne the data set.
56

China (Table 2.18 in Appendix). Table 2.2 shows that the greatest number of Korean
ﬁrms’ overseas affiliates is in the electronic and other electrical equipment and
components industry, and the second greatest share of the affiliates is in the apparel and

other ﬁnished products industry.8

Table 2.1: Locations of Afﬁliates9 and Equity Shares (%)

 

 

 

Location Observation Mean Median Min St. dev.
Asia 7,635 79.5 100.0 0.2 27.4
China 5,526 81.3 100.0 0.2 26.5
Japan 113 63.9 66.0 0.3 34.9
Indonesia 387 78.4 90.0 0.2 24.6
North America 671 75.3 100.0 0.2 35.5
US. 634 75.4 100.0 0.2 18.4
Europe 255 73.6 98.0 0.4 31.4
South America 241 88.5 100.0 1.7 22.9
Oceania 74 83.1 100.0 12.0 26.9
Central Asia 18 41.7 43.0 5.6 25.6
Affica 40 80.4 100.0 25.0 23.5
Total 8,934 79.2 100.0 0.2 28.2

 

This data set includes all Korean overseas investments from 1969 to the end of
March, 2003, and the initial observations in the data set number over 25,600. It is
obligatory for any agent (individual and / or ﬁrm) who (which) wants to invest overseas
and run a business to report to the Government (or government affiliated organization:

the Export-Import Bank of Korea). The original data set includes foreign investments by

 

a For Korean ﬁrms’ overseas joint ventures, see Table 2.19 in Appendix.
9 In this chapter, the term “Affiliates” in the title of all the tables denotes all affiliates in the
manufacturing industry acquired by Korean ﬁrms from 1973 to March 2003.

57

individuals as well as ﬁrms. We eliminate all individual investments from the data set
since foreign investments by individuals are not suitable for our analysis. This is
because most individual investors probably will not have advanced technologies which

could be appropriated by local companies in the host countries.

Table 2.2: Industries of Afﬁliates and Equity Shares (%)

 

 

Code Observation Mean Median Min St. dev.

20, 21 536 72.5 75.0 1.2 27.9
22 725 79.6 100.0 0.2 27.6
23 1,146 82.2 100.0 3.8 26.1
24 175 78.2 100.0 4.0 27.8
26 126 79.5 100.0 4.0 26.7
27 81 84.3 100.0 2.0 27.5
28 457 72.1 87.8 0.5 31.1
29 14 74.1 83.3 20.0 29.3
30 335 76.1 100.0 0.7 28.1
31 595 84.6 100.0 0.5 25.6
32 303 68.9 68.0 4.9 29.1
33 235 73.8 90.0 0.2 29.8
34 393 81.1 100.0 0.7 26.4
35 555 76.6 100.0 2.3 30.0
36 1,671 81.1 100.0 0.3 28.7
37 388 71.7 75.0 5.0 29.0
38 192 74.2 100.0 1.2 31.9

25, 39 992 85.7 100.0 0.2 24.8

 

Note: For explanation of industry code, refer to Table 2.16 in Appendix.

58

In concentrating on the manufacturing sector, foreign affiliates in non-
manufacturing sectors are leff out. Foreign direct investments in non-manufacturing
sector might not be directly related with the host country’s standards of intellectual
property rights protection. Furthermore, there were cases in which Korean multinational
ﬁrms had withdrawn their investment amounts, so that Korean firms did not have any
net investments for those affiliates. We also eliminate those cases from the original data
set. After polishing the data set, we have more than 8,900 overseas affiliates in the
manufacturing sector invested from 1973 to the end of March 2003, across 98 countries
(Table 2.17 in Appendix). Table 2.3 shows that overseas direct investment of Korean

ﬁrms has rapidly increased since the 19905. '0

Table 2.3: Year of Establishment and Equity Shares (%) of Afﬁliates

 

 

 

Year Observation Mean Median Min St. dev.
1973-1980 16 54.4 49.5 5.6 33.7
1981-1985 29 67.5 65.0 4.0 33.5
1986-1990 493 77.2 95.0 0.2 27.5
1991-1995 2,997 74.5 90.0 0.2 27.9
1996-2000 3,017 80.1 100.0 0.2 28.7
2001-2003 2,3 82 84.8 100.0 0.3 26.8

Total 8,934 79.2 100.0 0.2 28.2

 

Among 8,934 overseas affiliates, 5,200 affiliates are wholly-owned by Korean ﬁrms.

To test the relationship between intellectual property rights and equity structure by using

 

'0 The Korean Government has gradually liberalized its outward foreign direct investment since the ﬁrst
overseas investment. However, it was not until 1987 that a truly liberal policy on outward foreign direct
investment was adopted (Kim and Kang , 1997).

59

only partially-owned affiliates (pure joint ventures), we exclude wholly-owned (100%)
overseas affiliates, and arrive at 3,734 overseas pure joint ventures. With these pure
joint ventures, we can also test whether the degree of equity participation chosen by a
Korean multinational ﬁrm is affected by the host country’s intellectual property rights

protection. Table 2.4 shows that most joint ventures have been formed since the 1990’s.

Table 2.4: Year of Establishment and Equity Shares (%) of Afﬁliates (Pure Joint

 

 

 

Ventures)

Year Observation Mean Median Min Max St. dev.
1973-1980 13 43.9 49.0 5.6 99.8 27.9
1981-1985 16 41.1 42.5 4.0 90.0 20.9
1986-1990 285 60.5 55.0 0.2 99.7 25.4
1991-1995 1,563 51.1 50.0 0.2 99.9 18.8
1996-2000 1,168 48.4 50.0 0.2 99.7 22.3
2001-2003 689 47.5 50.0 0.3 99.9 22.9
1991-2003 3,420 49.5 50.0 0.2 99.9 21.0

Total 3,734 50.2 50.0 0.2 99.9 21.6

 

For the measurement of intellectual property rights of the host countries, we take the
popular index for the rights from Ginarte and Park (1997).” The Ginarte-Park index
takes into account ﬁve categories of patent laws: (1) extent of coverage, (2) membership
in international patent agreement, (3) provisions for loss of protection, (4) enforcement
mechanisms, and (5) duration of protection. Each of the categories is assigned a value

between 0 and 1, and the unweighed sum of these values constitutes the patent rights

index. The index (ipr) ranges from zero to ﬁve with the higher values indicating a

 

" In their paper, there is an index for 1990, but we cannot ﬁnd an updated index. We thank Park and
McCalman for sharing the extended index for 1995.
60

stronger level of protection.

To control other characteristics of the 98 host countries in the data set, we add
various control variables. It is possible that more human capital in the host country
(log hl) induces multinational ﬁrms in the source country to invest. In other words, the
more educated the labor force in the host country, the easier it might be for a
multinational ﬁrm to teach the techniques which are needed by the affiliates. Thus,
log hl might be positively related to equity participation. However, more human capital
in the host country might make it easier for the local companies to steal multinational
ﬁrms’ technologies. Hence, that variable may work both ways. Country-level
measurement for human capital abundance for the year 1988 is adopted from Hall and
Jones (1999).

To control the market size of a host country, we add the logarithm value of the host
country’s purchasing-power parity adjusted gross domestic product (lg dp) . We
multiply real12 gross domestic product per capita by a host country’s population. Each
host country’s real gross domestic product per capita of 2000 and population are
obtained from the Penn World Table.

In addition, we add telephone mainlines per 1000 inhabitants (phone) to control
the host country’s infrastructures. Although it is widely known that multinational ﬁrms
are inclined to invest in countries that have highly ordered infrastructures, it is not clear
whether a multinational ﬁrm increases its equity participation when the ﬁrm invests in a
host country with a well-equipped infrastructure. We take this variable13 from the
United Nations Conference on Trade And Development (UNCTAD). This variable is

scaled from 0 to 1 and the higher score of phone means that the host country has a

well-equipped infrastructure.

 

‘2 In the Penn World Table, “real” means “purchasing-power parity converted” instead of “in constant
rice”.
3 The telephone mainline is averaged from 1998 to 2000.

61

The measurement of a host country’s openness to foreign direct investment (fdi)
is available from Miles et al. (2003). This index takes into account factors such as
restrictions on foreign ownership of business, restrictions on industries and companies
open to foreign investors, and restrictions on repatriation of earnings. In addition, we
borrow an indicator of the share of black market economy from Miles et al. This index
(black) includes factors such as corruption, smuggling, and the underground provision
of labor and other services. It is known that a multinational ﬁrm which enters a country
with a large black market is likely to depend on its local partner to cope with obstacles
in the host country. Thus, the black market might be negatively related to the firm’s
equity participation. These two indices range from 0 to 5. The higher the score of fdi
(black) means that the less openness to foreign investors (the higher share of black
market, respectively) a host country enjoys.

The theoretical part of the previous chapter predicts that if the market becomes more
competitive, a multinational ﬁrm lowers its equity participation in overseas affiliate. To
test this relationship, we use the characteristics of the US. manufacturing industry. '4 We
use the logarithm value of the Herﬁndahl-Herschmann index15 (log(hhi)) taken from
US. Department of Commerce (2001) as a measure of market competition.'6 The
higher value represents less competition in the market.

To control for other potential determinants of internationalization, we add other
industry characteristics of the US. We add the logarithm value of the ratio of total
capital stock to total employment (log(k/l)), since the capital intensity of the industry

might affect the degree of equity participation. The logarithm value of human capital

 

'4 In section 4.4, we assume that all countries in the world (or all OECD countries) have the same market
structure to that in US. althouth this is a very strong assumption.
'5 The index is calculated by summing the squares of the individual company percentages of value added
for the 50 largest companies or the universe for 1997, whichever is lower.
‘6 The industry classiﬁcation for Herﬁndahl-Herschmann index follows the North American Industry
Classiﬁcation System (NAICS). Thus, we convert the NAIC classiﬁcation to Korean Standard Industrial
Classiﬁcation categories.

62

intensity of industry (log(h/l)) is added as a control variable since human capital
intensity of the production process might affect the integration decision. Human capital
intensity is measured as the ratio of non-production workers to production workers in a
given industry. In addition, the decision to integrate might be associated with the
importance of affiliate’s production in the overall value chain (Antras, 2003). To control
this possibility, we incorporate the logarithm value of the ratio of value added to total
sales (log(vad / sales)). We take the data for US. industry characteristics from the
NBER Manufacturing Industry Productivity Database.

We add the logarithm value of the average business-unit sales (log(scale)), since
the equity ownership decision may be affected by the size of the affiliate. To control the
importance of R&D in the production process, we also incorporate the logarithm value
of the ratio of R&D expenditures to sales (log(rnd / sales)). We take these two control
variables17 from Cohen and Klepper (1992).

To test the relationship between a Korean multinational ﬁrm’s equity participation
and the host country’s standards of intellectual property rights, we make the following
speciﬁcation

log(equityij) = 6] + 62iprj + Wj'tS’3 + (1’64 + 31.]. (2.1)
where subscript j represents host country j , log(equityij) is the logarithm value of
equity share (%) of the affiliate i which is located in country j. The Wj'. is the
vector of controls, d' is the vector of dummy variables, and al.]. is an error term. In

this speciﬁcation, we add the quadratic term of intellectual property rights (iprz) to
test whether McCalman’s results are valid in the production process. McCalman (2004)

showed that the relationship between the probability of foreign direct investment and

 

'7 Business-unit R&D expenditures and sales are averaged over the period 1974-1976 and 1975-1977,
respectively, to control for differences in the impact and timing of business cycle across industries.
63

intellectual property rights is U-shaped, based on the governance structure of the
operation of major Hollywood studios in 40 foreign markets in 1997.

In contrast to our focuses on the production process, he focused on the distribution
process of the movie industry. He demonstrated that increasing intellectual property
rights from a relatively low base tends to raise the attractiveness of licensing relative to
foreign direct investment. However, beyond a certain point, further increases in
intellectual property rights are associated with an increase in the likelihood of foreign
direct investment. The reason is that as the standards of intellectual property rights
protection are increased beyond said point, the proﬁts increased by the source country
ﬁrm become more important, and hence the optimal assignment of residual rights

moves from the host country ﬁrm to the source country ﬁrm (McCalman, 2004, p. 110).

Table 2.5: Descriptive Statistics for Intellectual Property Rights

 

 

Observation Mean St. dev. Min Max

log(equity) 8,934 4.25 0.60 -1.61 4.61

W 8,795 2.09 1.01 0 4.86

iprz 8,795 5.40 6.00 0 23.61

log hl 8,502 0.75 0.17 0.12 1.21

18 dp 8,443 7.84 1.34 1.43 9.19
phone 8,865 0.22 0.27 0.004 1
fdi 8,881 3.53 0.78 1 5
black 8,881 3.34 0.79 1 5

 

The vector Wj is included to control for other possible country-speciﬁc

determinants of equity participation and d' is incorporated to control the difference of

affiliate’s location,18 industry, and investment year. The theory expects a negative sign

of 62. To test a multinational ﬁrm’s preference for joint ownership when its affiliate is

in resources-based industries, we use both ordinary least squares method and probit
(logit) procedures. Tables 2.5'9 and 2.6 report descriptive statistics for all variables

included in the regressions to test the hypothesis.

Table 2.6: Descriptive Statistics for Intellectual Property Rights (Pure Joint

 

 

Ventures)
Observation Mean St. dev. Min Max
log(equity) 3,734 3.76 0.68 -1 .61 4.61
ipr 3,659 2.09 1.00 0 4.86
ipr2 3,659 5.41 5.88 0 23.61
loghl 3,582 0.74 0.17 0.12 1.21
lg dp 3,543 7.74 1.31 2.14 9.19
phone 3,713 0.22 0.27 0.004 1
fdi 3,719 3.48 0.77 1 5
black 3,719 3.37 0.82 1 5

 

To test the hypothesis that the more competitive the market is, a smaller equity share

a multinational ﬁrm holds, the following speciﬁcation is made

log(equityim) = a, + a, log(hhim ) + W129, + W554 + 5m (2.2)

 

‘8 Compared to the previous chapter; “Outsourcing, Joint Ventures, and Whole Ownership”, we do not
classify the host countries in the data set based on the origin of the country’s initial law as in La Porta et
al.’s paper (1998). We cannot categorize all 98 host countries in our data set into four different groups, as
in their paper, since they divided only 49 countries. Our data set has many Eastern European countries
and under-developed countries.

'9 There are some correlations among control variables to test the relationship between intellectual
property rights protection and the ﬁrm’s equity participation. However, the multicollinearity problem can
be mitigated by a huge number of observations (Wooldridge, 2000). In addition, in the following section,

we ﬁnd a robustly negative sign of 62 in speciﬁcation (1) which is the parameter of interest.
65

where subscripts m and j represent industry m and host country j, respectively.

In the speciﬁcation, log(equityim) is the logarithm value of equity share (%) of

affiliate i in industry m , log(hhim) is the logarithm value of industry m ’5
Herﬁndahl-Herschmann index, W' and W' are vector controls for host country

characteristics20 and industrial characteristics, respectively, and éim is an error term.

At this speciﬁcation, the theory expects a positive sign of 62. Descriptive statistics for

the variables for the regression are described in the Table 2.7.21

Table 2.7: Descriptive Statistics for Market Competition (All Afﬁliates)

 

 

Observation Mean St. dev. Min Max
log(equity) 8,934 4.25 0.60 -1.60 4.60
log(hhi) 8,917 4.49 0.77 2.14 6.68
log(k/I) 8,917 1.71 0.31 1.14 2.81
log(scale) 8,917 2.06 0.24 1.74 3.28
log(md / sales) 8,917 -2.06 0.42 -2.79 -1.41
log(h / I) 8,917 -0.47 0.20 -0.74 -0.02
log(vad / sales) 8,917 -0.30 0.06 -0.74 -0.17

 

4. Empirical Findings

First, we test the relationship between the intellectual property rights of the host
country and the Korean ﬁrm’s equity participation by using two different samples: a
sample of all overseas affiliates, and a sample of pure joint ventures. Second, we study

the U-shaped relationship between the intellectual property protection and ﬁrm’s equity

 

2° We also run regressions without controlling the country characteristics (see section 4.4).
2‘ Descriptive statistics for the affiliates in the OECD and the affiliates in the US. are shown in Tables
2.23 and 2.24 in Appendix, respectively.

66

participation. In addition, we test the relationship between investor protection and a
multinational ﬁrm’s equity ownership. Third, we examine the effects of resources-based
sector on the ﬁrm’s equity participation. Then, we study a multinational ﬁrm’s
preference for joint ownership. To test the relationships, we use ordinary least squares
method, and probit (logit) procedures. When we use a sample of pure joint ventures,
however, we run truncated regressions22 and ordinary least squares regressions. Finally,

we test the relationship between market competition and ﬁrm’s equity ownership.

4.1 Intellectual Property Rights and Equity Participation

(With All Afﬁliates)

We use a sample of all affiliates to test the relationship between the host country’s
intellectual property rights protection and the Korean ﬁrm’s equity participation. No
controls in the regression are included in column I of Table 2.8. The coefficient on
ipr is negative and signiﬁcantly different from zero at the 1% signiﬁcance level. It is
implied that a standard one step higher of the host country’s intellectual property rights
lowers the Korean ﬁrm’s equity participation by about 9.6%. Column 11 includes the

host country’s human capital abundance in the regression. This inclusion does not affect

the signiﬁcance of the ipr. However, this leads to an increase of the absolute value of

62 estimate. The coefficient on log hl is positive and statistically signiﬁcant at the

1% signiﬁcance level. This implies that the more abundant human capital a host country
has, the greater equity share a Korean ﬁrm is inclined to hold.
In column III, the measurement for the host country’s market size is added in the

regression. The coefficient on 1g dp is negative and statistically signiﬁcant. This result

 

22 A truncated regression model arises when we exclude, on the basis of the dependent variable, a subset
of the population in sampling scheme. In other words, we do not have a random sample from the
underlying population, but we know the rule that was used to include units in the sample (Wooldridge,
2000).

67

might be interpreted as contradicting the recognition that the host country’s market size
attracts foreign direct investment. However, the dependent variable is not the quantity of
foreign direct investment; rather it is the form of foreign direct investment.23 In order to
more accurately test the relationship between market size and multinational ﬁrm’s
equity ownership, the measurement for market size should be reﬁned. The purchasing-
power parity adjusted gross national product of the host country might not be sufficient
to control the host country’s market potential since the multinational ﬁrms can export to
nearby countries when the ﬁrms use its affiliate as both an overseas production plant (or
an overseas intermediate-input supplier) and an export platform.24 In addition, it is
possible that a Korean multinational ﬁrm has a smaller equity share of local company in
the larger host country’s economy since a larger economy might have bigger local

25 When we add all control variables in column VI, lgdp becomes

companies.
insigniﬁcant.

In column IV, the coefficient on phone is negative but not statistically signiﬁcant.
However, when we add all control variables in column VI, the coefficient becomes
signiﬁcant. It is interpreted that if a host country has a well-equipped infrastructure,
then a Korean multinational ﬁrm acquires a lower equity share of its affiliate. This result
might be interpreted as the following; if a local company is located in a country with
favorable environments for multinational ﬁrms, then it has more bargaining power than
local companies in other countries. The ipr is still signiﬁcantly negative after we
control the host country’s degree of infrastructure.

In column V, we include the host country’s openness to foreign direct investment.

The coefficient on fdi is positive, but not signiﬁcant. This variable does not affect the

 

2’ However, a multinational ﬁrm is likely to hold a greater share of affiliate’s equity if a larger equity
share induces the ﬁrm to acquire a larger fi'action of its afﬁliate’s proﬁts.
2" For more detail, see Head and Mayer (2004), and Neary (2002).
25 La Porta et a1. (1998) found that larger economies have a lower ownership concentration by examining
domestic ﬁrms’ choices in 49 countries.

68

qualitative results on multinational ﬁrm’s equity participation. The coefficient on

intellectual property rights remains signiﬁcant at the 1% signiﬁcance level.

Table 2.8: Intellectual Property Rights and Equity Ownership Structures

 

 

 

I II III IV V VI
Dep. Variable= log(equity)
ipr -.112*“ -.l36"” -.124"* -.124*** -.ll8"”
(.013) (.018) (.020) (.019) (.019)
log hl 270"" 350*" .401 *" 393‘" 259*”
(.075) (.085) (.091) (.092) (.100)
lg dp -.017" -.017” -.020** -.014
(.008) (.009) . (.010) (.010)
phone -.099 -.078 -.4 l 4*”
(.062) (.077) (.125)
fdi .009 -.007
(.020) (.021)
black -.103***
(.030)
dummies for location, year, and industry.
Obser. 8,394
R2 .06 .06 .06 .06 .06

 

Note: Standard errors in parenthesis.
(", and *** are 5%, and 1% signiﬁcance levels, respectively.)

In column VI, the indicator of the host country’s share of the black market in the

economy is included in the regression. The share of the black market in the host

country’s economy has an important effect on the Korean ﬁrm’s equity participation. It

69

is widely recognized that ﬁrms which enter countries with a large black market are more
likely to contend with obstacles in their businesses. In such environments, a local
partner can more effectively deal with the problems related to the host country (Asiedu
and Esfahani, 2001). As a result, the black market is expected to be negatively related to
the ﬁrms’ equity participation.26 This is conﬁrmed strongly by the regression. The

coefficient on black is negative at the 1% signiﬁcance level.

(With Pure Joint Ventures)

There is a possibility that some multinational ﬁrms plan their entry modes in two
stages. In the ﬁrst stage, a multinational ﬁrm chooses between whole ownership and a
joint venture. In the second stage, the multinational ﬁrm should divide the share of its
afliliate’s equity with its local partner in the host country if the ﬁrm chose a joint
venture as its entry mode in the ﬁrst stage. To consider this possibility, we make a
sample of pure joint ventures by excluding wholly-owned affiliates in the data set and
test the relationship between the host country’s intellectual property rights protection
and a multinational ﬁrm’s equity participation.

We test whether the intercept and all slopes are the same between two samples: a
sample consisting of all affiliates, and another sample consisting of purely joint
ventures.27 The F test shows that the p -value is zero, which causes us to soundly
reject the null hypothesis that the intercept and all slopes are the same among two
groups. Thus, we run regressions with pure joint ventures to test the relationship
between the host country’s property rights protection and a multinational ﬁrm’s equity

participation in its affiliate.

 

2" We might conclude that a large black market indicates weak legal institutions, implying that the risk of
technology leakage would be enhanced (with no recourse to courts). However, the ipr is thought to

represent some characteristics of the host country’s legal institutions. In addition, we analyze the effects
of the host country’s legal institutions on the multinational fum’s equity participation more in detail in
section 4.2.
27 We test whether the intercept and all slopes are the same across the two samples by using dummy
variables.

70

We run regressions with a sample of partially-owned affiliates (pure joint ventures)
by excluding wholly-owned (100%) affiliates to test aforementioned hypothesis. The
regression algorithm using the method of ordinary least squares might lead to
potentially biased estimates due to a truncated dependent variable. Thus, we run
truncated regressions to estimate variable coefficients (Table 2.9). However, we also
convey the results obtained by the method of ordinary least squares regressions in Table
2.20 in Appendix.

Table 2.9 shows control variables’ estimates obtained by truncated regressions. The
estimate in column I is negative and statistically signiﬁcant at the 1% signiﬁcance
level as in Table 2.8. When only partially-owned affiliates are considered, a standard
one step higher of the host country’s intellectual property rights lowers the Korean

ﬁrrn’s equity participation by about 14.9%. In column II , human Capital abundance is

added in the regression. The coefﬁcient on log hl is positive but not statistically
signiﬁcant. However, log hl becomes signiﬁcant when we add all control variables.

In column 111, the lg dp is signiﬁcantly negative. However, the lgdp might not
represent the real market size of the host country since this control variable does not
represent nearby markets of a host country, as already mentioned. In column IV, the
measurement for host country’s infrastructures is included, and the coefficient is
positive, but not signiﬁcant. However, when we add all control variables, it becomes
statistically signiﬁcant (column VI). In column V, fdi is not signiﬁcant. The
coefficient on the black market is signiﬁcantly positive at the 5% level in column VI.
The consecutive inclusions of control variables do not overturn the qualitative results
for the negative relationship between the host country’s standards of intellectual

property rights protection and the Korean ﬁrm’s equity participation.
We ﬁnd that ipr is negative and statistically signiﬁcant at the 1% level regardless

of regression methods (Tables 2.9 and 2.20 in Appendix). The absolute values of ipr
71

in the ordinary least squares method (Table 2.20 in Appendix) are similar to those in the
truncated regressions (Table 2.9). There is little difference in qualitative results between
Tables 2.9 and 2.20. Thus, we can conclude that a negative relationship between the
host country’s standards of intellectual property rights protection and a Korean

multinational ﬁrm’s equity participation in its affiliate is very robust.

Table 2.9: Intellectual Property Rights and Equity Ownership Structures

 

 

 

(Pure Joint Ventures)
I II III IV V VI
Truncated Regressions. Dep. Variable= log(equity)
ipr -.l49"* -.155"* -.220"”""l -.223“* -.223""‘ -.228"”'“‘I
(.019) (.021) (.027) (.030) (.030) ( .030)
log h] .089 360*" .348" .348M .452‘”
(.110) (.131) (.138) (.140) (.149)
lgdp -.001**"‘ -.001*" -.001"* -.001"'"‘"
(.000) (.000) (.000) (.000)
phone .027 .027 .3 50"
(.098) (.116) (.193)
fdi ' -.000 .010
(.030) (.030)
black .099"
(.047)
dummies for location, year, and industry.
Obser. 3,519
Log-Like“. 282.55 282.91 289.53 289.56 289.68 292.21

 

Note: Standard errors in parenthesis.

(‘, ", and *" are 10%, 5%, and 1% signiﬁcance levels, respectively.)

72

4.2 Quadratic Relationship and Investor Protection

We test whether U-shaped relationship between the intellectual property rights and
the multinational ﬁrms’ equity participations is still valid in the production process. As
mentioned, McCalman (2004) showed that the relationship between the probability of
foreign direct investment and intellectual property rights is U-shaped by studying the
distribution process of the movie industry. We run regressions with a sample of all
affiliates (column I ) as well as a sample of pure joint ventures by excluding wholly-
owned affiliates (column H and III).

Columns I and H of Table 2.10 are estimated by ordinary least squares method,
but column III is obtained by truncated regressions. As McCalman showed that the U-
shaped relationship exists in the distribution process, so columns I and III conﬁrms
that there is still an U-shaped relationship in the production process. This implies that
the standards of intellectual property rights have a diminishing effect on a multinational
ﬁrm’s equity participation in its affiliate. Since the measure of intellectual property
rights ranges from 0 to 5, the minimum level occurs within this range. Solving for the
turning point for column I (for column III) reveals that it occurs at an intellectual
property rights’ level of approximately 4.61 (4.89, respectively). Before this point, the
level of intellectual property rights has a negative effect on the multinational ﬁrm’s
equity participation in its overseas affiliate; after this point, the level of intellectual
property rights has a positive effect on the equity participation.

However, when we consider whole affiliates (column I ), only the US. has a
greater index (4.86) than the turning point (4.61). The quadratic term is positive but not
statistically signiﬁcant in column II. The parabolic shape of the relationship between
the host country’s intellectual property rights and a multinational ﬁrm’s equity

participation is not supported in regressions of ordina1y least squares with a sample of

73

pure joint ventures.

Table 2.10: Quadratic Relationship
I (All Affiliates) H (Joint Ventures) III (Joint Ventures)

Dep. Var.= log(equity)

 

 

Ordinary Least Squares Truncated Regression
ipr -0.277*** -0.460*** -0.499***
(0.086) (0.153) (0.156)
iprz 0.030“ 0.041 0.051 *
(0.016) (0.028) (0.029)
log h! 0.308“ 0498*" 0545*"
(0.103) (0.155) (0.158)
lg dp -0.027** 0075*" 0001*"
(0.012) (0.023) (0.000)
Phone -0.446*** 0.419" 0.215
(0.126) (0.207) (0.207)
fdi 0.001 0.039 0.028
(0.021) (0.035) (0.032)
black -0.111*** 0.109" 0.065
(0.03 0) (0.049) (0.051)
dummies for location, year, and industry.
Observation 8,3 94 3,5 19 3,5 19
R2 0.06 0.14 -
Log-Likeli. - - 290.24

 

Note: Standard errors in parenthesis.
C“, ", and **"‘ are 10%, 5%, and 1% signiﬁcance levels, respectively.)

74

La Porta et al. (1998) tested the hypothesis that ﬁrms in countries with poor investor
protection have more concentrated ownership of their shares. They emphasized that
large, or even dominant, shareholders who monitor the managers might need to own
more capital to exercise their control rights and, thus, to avoid being expropriated by the
managers. This would be especially true when there are some legal or economic reasons
for large shareholders to own signiﬁcant cash ﬂow rights as well as votes. With poor
investor protection, ownership concentration becomes a substitute for legal protection,
because only large shareholders can hope to receive a return on their investment (p.
1145).

They classiﬁed 49 countries into four groups28 based on the origin of the country’s
initial law: such as common-law countries, French-civiI-law countries, Gerrnan-civil-
law countries, and Scandinavian-civil-law countries. They found that common-law
countries generally have the strongest, and F rench-civil-law countries have the weakest,
legal protections for investors. In addition, they showed that a concentration of
ownership of shares in the public ﬁrms is negatively related to investor protections.
They explained that the results of uncommonly high ownership concentration in French-
civil-law countries are suggestive that concentration of ownership is an adaptation to
poor legal protection (p. 1148).

Following their categorization of countries, we classify the host countries in the data
set into ﬁve groups. Since their paper covered only 49 countries, the majority of host
countries belong to the other country group (Table 2.11). The other country group
consists of underdeveloped countries, Eastem-Europe countries, and socialist countries.
Thus, we suspect that countries in the other group have much weaker legal protections

for investors than French-civil-law type countries. However, the common-law countries

 

28 For more detail, see the previous chapter; “Outsourcing, Joint Ventures, and Whole Ownership”.
75

 

have the strongest legal protections for investors, as La Porta et al. showed.

Table 2.11: Countries by Legal Origin

 

Origins

Countries

 

Common-law

Australia, Canada, Hong Kong, India, Ireland, Malaysia, Singapore,
South Africa, Sri Lanka, Thailand, U.K., U.S., Kenya, New Zealand,

Pakistan.

 

F rench-civil-law

Argentina, Belgium, Brazil, Chile, France, Indonesia, Italy, Mexico,
Netherlands, Peru, Philippines, Portugal, Spain, Turkey, Columbia,

Ecuador, Jordan, Venezuela, Uruguay, Egypt.

 

Gerrnan-civil-

Austria, Germany, Japan, Switzerland, Taiwan.

 

 

law

Scandinavian- Sweden, Denmark, Finland.

civil-law

Other Bangladesh, China, Vietnam, Uzbekistan, Czech Rep., Hungary,

Poland, Romania, Russia, Panama, Virgin Islands, Guam, Morocco,
Sudan, Tunisia, Nepal, Cameroon, Central African Republic, Congo,
Angola, Ghana, Mauritius, Senegal, Togo, Tanzania, Bolivia, Iran,
Paraguay, Costa Rica, El Salvador, Guatemala, Honduras, Bahamas,
Nicaragua, Jamaica, Dominican Rep., Oman, Qatar, Syria, Fiji,
Saudi Arabia, United Arab Emirates, Myanmar, Cambodia, Laos,
Kyrgyzstan, Tajikistan, Mongolia, Malta, Lithuania, Samoa,
Slovenia, Solomon Islands, Papua New Guinea, Northern Mariana

Islands.

 

By using the Korean ﬁrm-level data set, we test the hypothesis that a multinational

76

ﬁrm holds a greater equity share of its affiliate when the affiliate is located in countries
with poor investor protection. Although ipr concentrates on the standards for foreign
countries’ intellectual property rights, the measure of the legal system focuses on lawful
protections for investors.” 3°

To test the hypothesis, we make four dummy variables; french which represents
the French-civil-law countries, german which stands for the German-civil-law
countries, scandinavian which represents Scandinavian-civil-law countries, and
other which stands for countries in other country group. We choose the common-law
countries as a base group to avoid a dummy variable trap (Wooldridge, 2000). Putting
these four dummy variables into speciﬁcation (1) gives the below result. The estimates
on the four dummy variables measure the proportionate differences in Korean ﬁrms’
equity participations against equity participations in afﬁliates in common-law countries.

Table 2.12 shows that the estimate on french shows that a Korean multinational
ﬁrm acquires about 18.2% more equity share of affiliate in French-civil-law countries
than the equity share of affiliate in common-law countries, holding levels of human
capital abundance, gross domestic product, telephone mainlines, openness, and share of
black market ﬁxed. In addition, a Korean ﬁrm holds around 28.1% more equity share of
an afﬁliate in countries in other country group than in common-law countries. Those
coefficients are statistically signiﬁcant at the 1% level.

However, the coefficient on german is negative and statistically signiﬁcant at the
10% level (column I ). And the coefficient on scandinavian is negative but not

statistically different ffom zero (columns I , II, and III). Columns 11 and 111 show

 

29 Thus, the results in this section can strengthen our ﬁndings of a strong negative relationship between
host country’s standards of intellectual property rights protection and the multinational ﬁrrn’s equity
participation.

3° When we run regressions either including ipr or without ipr , we ﬁnd that there are no differences in
qualitative results between the two regressions. Hence, this measure of legal systems can explain other
aspects of multinational ﬁrm’s choices which ipr does not.

77

that the qualitative results do not change when we run a regression by using a sample of
pure joint ventures except german . The coefficient on german is positive and
statistically signiﬁcant at the 10% level in column H , but not signiﬁcant in column III.
We ﬁnd from the results of Korean ﬁrm-level data set that high concentrated ownership

also results from weak protections of investors as La Porta et al. showed.

Table 2.12: Legal Origins and Equity Ownership Structures

 

I (All Affiliates) II (Joint Ventures) III (Joint Ventures)

 

Dep. Var.= log(equity)

 

 

 

Ordinary Least Squares Truncated Regression
french 0182*" 0428*" 0390*"
(0.03 8) (0.058) (0.085)
german -0.088* 0.155* 3 0.065
(0.052) (0.080) (0.083)
scandinavian -0.206 -0.501 -0.573
(0.342) (0.649) (0.643)
other 0.281 *** 0405*“ 0.412***
(0.032) (0.052) (0.050)
dummies for year, and industry.
W ' for host country characteristics.
Observation 8,3 94 3,519 3,519
R2 0.06 0.13 -
Log-Likeli. - - 277.85

 

Note: Standard errors in parenthesis.

(‘, and **"‘ are 10%, and 1% signiﬁcance levels, respectively.)

78

4.3 Resources-Based Sector and Preference for Joint Ownership

The local companies in the host country have an advantage in accessing supplies of
agricultural or mineral resources, and the quality of these resources might be very
important to the multinational ﬁrm. Although a multinational ﬁrm prefers full control
over its affiliate to exploit ﬁrm-speciﬁc assets, the ﬁrm tends to share control rights on
its affiliate with a local partner in the host country since bargaining power of the local
partner in resources-based industries might be much higher than that of the
multinational ﬁrm (Kobrin, 1987). There is a different explanation for this inclination
for joint ownership. These resources-based industries are least likely to depend on
complicated R&D capabilities. The relative experience of local companies would be
greater in such basic industries than more advanced sectors. This relative experience
argument expects the higher occurrence of joint ownership in resources-based industries
(Gomes-Casseres, 1989).

To test whether a multinational ﬁrm prefers shared control rights when it invests in
resources-based sector, we divide all industries into two groups to examine whether the
importance of natural resources in the industry affects the ﬁrm’s preference for
ownership. Gomes-Casseres (1990) introduced the dummy variable for resources-based
industries which include food and beverages (SIC 20), tobacco (SIC 21), textile mills
(SIC 22), wood products excluding furniture (SIC 24), pulp and paper (SIC 26),
petroleum (SIC 29), rubber (SIC 30), leather (SIC 31), stone and glass (SIC 32), and
primary metals (SIC 33). We use a dummy for resources-based industries (resource)
and test whether a multinational ﬁrm prefers shared ownership to whole ownership

when its affiliate is in resources-based sector.

79

Table 2.13: Resources-Based Sector

 

I (All Affiliates) II (Joint Ventures) III (Joint Ventures)

 

Dep. Var.= log(equity)

 

 

Ordinary Least Squares Truncated Regression
ipr -0.123*** -0.239*** -0.227***
(0.019) (0.032) (0.03 0)
108 hi 0.283M 0443*" 0.458"
(0.101) (0.149) (0.149)
lgdp -0013 -0.049*** 0001*"
(0.010) (0.017) (0.000)
Phone -0.439*** 0.480M 0365*
(0.126) (0.203) (0.193)
fdi -0.014 0.021 0.011
(0.021) (0.032) (0.030)
black -0.105*** 0130*" 0.106"
(0.030) (0.048) (0.047)
resource -0.030** 0.003 0.003
(0.014) (0.022) (0.022)
dummies for location, and year.
Observation 8,3 94 3,519 3,519
R?- 0.06 0.14 -
Log-Likeli. - - 281.06

 

Note: Standard errors in parenthesis.

(*, ", and ‘** are 10%, 5%, and 1% signiﬁcance levels, respectively.)

First, we run least squares regressions and truncated regressions to study the effects

80

of affiliates in the resource-based sector on a multinational ﬁrm’s equity participation.
Again, we use two different samples, a sample of all affiliates (column I in Table
2.13) and a sample of pure joint ventures (columns H and III). Table 2.13 shows that
the coefficient on resource is negative and statistically signiﬁcant at the 5% level in
column I . This implies that a multinational ﬁrm holds a smaller equity share when its
overseas affiliate is in the resources-based sector, as we expected. However, when we
run regressions with a sample of pure joint ventures, resource is positive but not
statistically signiﬁcant in columns II and III.

3‘ to examine whether a

Second, we employ probit and logit procedures
multinational frrm tends to share control rights on its affiliate with a local partner when
the affiliate is in the resources-based sector. To study the multinational ﬁrrn’s preference
for joint ownership, we do not use actual ownership level because the difference, say,
between 100% foreign ownership and 80% is likely to be perceived differently by ﬁrms
than that between 80% and 60% (Gomes-Casseres, 1990). We deﬁne the dependent
variable as a dummy variable. Following Franko (1971), Stopford & Wells (1972),
Gatignon & Anderson (1988), Gomes-Casseres32 (1989), and Hennart (1991), shared

ownership is deﬁned here as any affiliate where the Korean multinational ﬁrm owned

less than 95% of the equity. That is, the dummy variable33 for the dependent variable

 

3‘ We can consider a class of binary response models of the form P(y=1/x)=

G(ﬂ0 + ,lel + ...+ ,kak ) = G(ﬂO + xﬂ) , where G is a function taking on values strictly between zero
and one: 0 < G(z) < l , for all real numbers 2 . In the probit model, G is the standard normal

Z
cumulative distribution function, G(z) = I¢(v)dv (where ¢(z) = (27r)_1/2 exp(—z2 /2) ). However,

.w
in logit model, G is the logistic function, G(z) = exp(z)/[l +exp(z)]. To make the logit and probit

slope estimates comparable, we can either multiply the probit estimates by 1.6, or we can multiply the
logit estimates by 0.625. (Wooldridge, 2000).

32 Gomes-Casseres mentioned that Franko, and Stopford & Wells also deﬁned the dependent variable in
the same way. F ranko, Lawrence G. (1971) Joint Venture Survival in Multinational Corporations, New
York: Praeger. Stopford, John M., and Louis T. Wells Jr. (1972) Managing the Multinational Enterprise,
New York: Basic Books. Hennart, Jean-Francois (1991) “The Transaction Costs Theory of Joint Ventures:
An Empirical Study of Japanese Subsidiaries in the United States,” Management Science. 37(4), 483-497.
33 The mean of the dummy variable is 0.598 and the standard deviation is 0.490.

81

 

equals 1 if a Korean multinational ﬁrm owns greater than or equal to 95% of the
affiliate’s equity for the probit (column I in Table 2.14) and logit (column II)
procedures.

Table 2.14 shows that ipr is negative and statistically signiﬁcant at the 1%
signiﬁcance level in columns I and II. This implies that if ﬁrm’s affiliate is located
in a country with high standards of intellectual property rights protection, then the
Korean multinational ﬁrm prefers shared control rights. The coefficients on resource
are signiﬁcantly negative at the 1% level. This implies that there is a strong positive
relationship between affiliate in the resources-based sector and the probability of joint
ownership between a Korean multinational ﬁrm and its local partner in the host country.

The coefficients on black are negative and statistically signiﬁcant at the 1% level.
Korean multinational ﬁrms prefer shared ownership for their affiliates that are located in
countries with higher shares of the black market. In Table 2.14, we cannot ﬁnd a strong
relationship between Korean ﬁrm’s preference for joint ownership and the host
country’s market size. However, a Korean ﬁrm tends to share control rights when its
affiliate is in country with a well-equipped infrastructure.

This signiﬁcantly negative coefficient on phone might mean that a local company
in a host country with a good infrastructure has high bargaining power since the quality
of an economy’s infrastructure is an important element in investment decisions for
foreign investors.34 The coefficients on loghl and fdi are not statistically different

from zero. That is, we cannot ﬁnd any strong effects of log hl and fdi on the

multinational ﬁrm’s preference for shared ownership.

 

3’ Gomes-Casseres (1989) assumed that as the size of a country’s industrial sector grows, local companies
acquire commercial experience that may be valuable to a multinational ﬁrm. Thus, the beneﬁts of joint
ownership should increase with the size of the host county’s industrial sector. He found that the size of a
host country’s industrial sector had a strong positive effect on the probability of shared ownership. We can
speculate that the grth of industrial sector is closely related to that of infrastructure.

82

Table 2.14: Preference for Joint Ownership

 

 

 

I (Probit) H (Logit)
Dep. Var.=] if Equity Share 2 95%
ipr -0.130*** -0.216***
(0.043) (0.072)
log k] 0.283 0.473
(0.216) (0.356)
lg dp -0.016 -0.028
(0.023) (0.039)
phone -1 .301 *** -2.174***
(0.271) (0.447)
fdi -0.032 -0.055
(0.045) (0.075)
black -0.376*** -0.631***
(0.065) (0.108)
resource -0.l35*** -0.217"”MI
(0.030) (0.049)
dummies for location, and year.
Observation 8,394
Log-Likelihood -531.08 -532.88
Pseudo R2 0.05 0.05

 

Note: Standard errors in parenthesis. ("* is 1% signiﬁcance level.)

83

4.4 Market Competition and Equity Structures

Blomstrbm and Kokko (1997) emphasized there is one central problem in studies of
foreign direct investment. The problem is to discover whether a multinational ﬁrm’s
entry and presence explains industry structure of the host country, or whether industry
structure determines if multinational ﬁrms enter (p. 29). They emphasized that the
dynamic aspects of multinational ﬁrms and competition in host country markets are not
well researched.

Theoretically, Neary (2002) showed that increasing competition leads multinational
ﬁrms to withdraw from the market. In this section, we test the relationship between the
multinational frrm’s equity participation in its affiliate and market structure.35 We run
regressions with different observations in the data set, with all affiliates. in the world,
with affiliates located in the OECD countries, and with affiliates in the US. To run
regressions, we assume that the market competition and industry characteristics are
identical for all the countries in the world, or for all OECD countries.36 Tables 2.21 and
2.22 in Appendix describe the affiliates in the OECD and US. for each 2-digit
manufacturing industry, respectively.

Table 2.15 shows the relationship between market competition and ﬁrm’s equity
ownership. In column I , we ﬁnd a simple relationship between equity ownership
(log(equity)) and the Herﬁndahl-Herschmann index (log(hhi)). The coefficient on
(log(hhi)) is negative and statistically signiﬁcant at the 1% signiﬁcance level. In
column I], the ratio of capital stock to employment is added. The coefficient on

(log(hhi)) is still statistically signiﬁcant at the 10% level. The inclusion of average

 

’5 Neary (2002) analyzed horizontal foreign direct investment. However, the previous chapter,
“Outsourcing, Joint Ventures, and Whole Ownership”, focused on the vertical relationship and predicted
that if the market becomes more competitive, then a multinational ﬁrm decreases its equity ownership of
its affiliate.
36 As we already mentioned, this is a very strong assumption. We cannot ﬁnd appropriate industry
characteristics for all countries in the world, or OECD countries.

84

business-unit sales in column HI makes the Herﬁndahl-Herschmann index positive and

insigniﬁcant. The inclusion of the ratio of R&D expenditures to sales in column IV

does not change the qualitative results.

Table 2.15: Market Competition and Equity Ownership (All Affiliates)

 

 

 

I H HI IV V VI
Dep. Variable= log(equity)
log(hhi) -.022** * -.015* .007 .009 .008 .010
(.008) (.008) (.010) (.011) (.011) (.011)
log(k /1) -.172*** -.103*** -.081* -.085** -.096**
(.021) (.029) (.041) (.042) (.043)
log(scale) -.150*** -.162*** -.l65*** -.194***
(.045) (.048) (.048) (.058)
log(rnd / sales) -.017 -.046 -.043
(.023) (.032) (.033)
log(h/l) .077 .119
(.060) (.076)
log(vad / sales) -.170
(.193)
dummies for location, and year.
W' For host country characteristics.
Obser. 8,393
R2 0.04 0.05 0.05 0.05 0.05 0.05

 

Note: Standard errors in parenthesis.

(*, *"', and *** are 10%, 5%, and 1% signiﬁcance levels, respectively.)

When we incorporate all the control variables for industry characteristics, the

85

 

coefficient on (log(hhi)) is positive but not signiﬁcant. As a result, we cannot ﬁnd a
strong negative relationship between market competition and multinational ﬁrms’ equity
participation. After running regressions” using all affiliates (Table 2.25 in Appendix),
affiliates in the OECD countries (Table 2.26 in Appendix), or the affiliates in the US.
(Table 2.27 in Appendix), we still cannot ﬁnd a signiﬁcantly negative relationship
between market competition and a multinational firm’s equity participation. To test
more accurately the relationship between market competition and equity structure, we
might need more precise information on market structures of host countries.

However, some interesting results found include that there is a negative relationship
between the Korean ﬁrm’s equity ownership of affiliate and the industry’s ratio of
capital stock to employment. It means that a Korean ﬁrm holds a smaller equity share
when the ﬁrm’s affiliates are in a high capital ratio industry. The coefficient on
log(scale) is also negative and statistically signiﬁcant at the 1% signiﬁcance level. It
might be interpreted that if a Korean firm invests in a larger affiliate, it tends to hold a

smaller equity share due to the budget constraint and / or risk aversion.

5. Concluding Remarks

There are many factors which affect multinational fums’ choice of ownership
structures. We take into account several very important factors of the multinational
ﬁrms: the host country’s standards of intellectual property rights protection, resources-
based industries, and market competition. Most theoretical papers expect that
multinational ﬁrms prefer wholly-owned affiliates to licensing or joint ownership when
the ﬁrms have a high risk of technology leakage.

Using ﬁrm-level data, we ﬁnd a strong negative relationship between host country’s

 

’7 In all these regressions, we do not control host country’s characteristics.
86

 

standards of intellectual property rights and the multinational ﬁrm’s equity participation
in its affiliate. The evidence is taken from the two regressions, observations from all
affiliates and partially-owned affiliates. Keeping the other factors constant, host
country’s weak standards of intellectual property rights induce a multinational firm to
hold a greater equity share of its affiliate. In addition, the U-shaped relationship
between intellectual property rights protection and a multinational ﬁrm’s equity
participation is supported in regressions with a sample of all overseas affiliates.
However, it is strongly supported that high concentrated ownership also results from
weak protections of investors as La Porta et al. showed.

We ﬁnd that there is a strong relationship between multinational ﬁrrn’s preference
for joint ownership and its affiliate in the resources-based industries. Since the quality
of the resources might be important to a multinational firm and a local company in the
host country has an advantage in accessing a supply of resources, the multinational ﬁrm
is inclined to share the control rights over its affiliate with a local partner. However, we
cannot ﬁnd any deﬁnite evidence that a multinational ﬁrm lowers its equity share of
overseas affiliate when the host country market becomes more competitive. One
interesting result is that a multinational ﬁrm tends to prefer shared ownership when its

affiliate is located in a country with a higher share of the black market.

87

APPENDIX

Table 2.16: List of Industries

 

 

Code” Description
20 Food and kindred products
21 Tobacco products
22 Textile mill products
23 Apparel and other ﬁnished products made from fabrics and similar materials
24 Lumber and wood products, except furniture
26 Paper and allied products
27 Printing, publishing, and allied industries
28 Chemicals and allied products
29 Petroleum reﬁning and related industries
30 Rubber and miscellaneous plastic products
31 Leather and leather products
32 Stone, clay, glass, and concrete products
33 Primary metal products
34 Fabricated metal products, except machinery and transportation equipment
35 Industrial and commercial machinery and computer equipment
36 Electronic and other electrical equipment and components, except computer
equipment
37 Transportation equipment
38 Measuring, analyzing and controlling instrument; photographic, medical and
optical goods; watches and clocks
25, 39 Furniture and ﬁxture, miscellaneous manufacturing industries

 

Note: 1) Standard Industrial Classiﬁcation 1987.

88

Table 2.17: Countries by Income Level"

 

Income Level

Countries

 

~ 1000

Togo, Tanzania.

 

1 000~5 000

Morocco, Egypt, Cameroon, Congo, Ghana, Kenya, Senegal,
Bolivia, Ecuador, Paraguay, Peru, El Salvador, Guatemala,
Honduras, Nicaragua, Jamaica, Jordan, Syria, Bangladesh, Sri
Lanka, India, Indonesia, Nepal, Pakistan, Philippines, China,

Kyrgyzstan, Tajikistan.

 

5000~1 0000

South Africa, Tunisia, Brazil, Colombia, Mexico, Venezuela,
Costa Rica, Dominica, Panama, Iran, Turkey, Malaysia, Poland,

Thailand, Lithuania, Romania, Russia.

 

1 0000~20000

Mauritius, Argentina, Chile, Uruguay, Portugal, Spain, Czech

Republic, Hungary, Slovenia, [Korea (ROK)].

 

20000~

Canada, U.S., Japan, Hong Kong, Singapore, Belgium, France,
Denmark, Germany, Ireland, Italy, Netherlands, U.K., Austria,

Finland, Sweden, Switzerland, Australia, New Zealand.

 

Others”

Sudan, Central Aﬁican Republic, Angola, Bahamas, Guam,
Virgin Island, Oman, Qatar, Saudi Arabia, United Arab Emirates,
Myanmar, Cambodia, Laos, Taiwan, Uzbekistan, Mongolia,
Vietnam, Malta, Fiji, Solomon Islands, Papua New Guinea,

Samoa, Northern Mariana Islands.

 

Note: 1) Purchasing-power parity converted gross domestic product per capita,
Penn World Table (2000).
2) Data is not available.

89

Table 2.18: Locations of Afﬁliates and Equity Shares (%) (Pure Joint Ventures)

 

 

 

Location Observation Mean Median Min Max St. dev.
Asia 3,213 51.3 50.0 0.2 99.9 20.3
China 2,097 50.6 50.0 0.2 98.9 18.4
Japan 71 42.5 50.0 0.3 95.0 26.5
Indonesia 267 68.6 70.0 0.2 99.9 24.0
North America 263 36.9 39.2 0.2 99.9 28.3
US. 243 35.8 33.3 0.2 99.7 27.9
Europe 129 47.8 50.0 0.4 99.8 24.7
South America 64 56.7 50.0 1.7 99.5 24.7
Oceania 30 58.3 50.0 12.0 99.0 27.4
Central Asia 16 34.5 36.0 5.6 50.0 15.3
Africa 19 58.8 55.0 25.0 89.0 16.1
Total 3,734 50.2 50.0 0.2 99.9 21.6

 

90

Table 2.19: Industries of Afﬁliates and Equity Shares (%) (Pure Joint Ventures)

 

 

Code Observation Mean Median Min Max St. dev.

20, 21 309 52.2 50.0 1.2 97.2 19.5
22 307 51.9 50.0 0.2 99.0 21.7
23 427 52.1 50.0 3.8 99.3 19.7
24 81 53.0 50.0 4.0 98.6 21.8
26 56 53.7 52.3 4.0 90.0 20.4
27 28 54.6 53.4 2.0 99.7 29.1
28 242 47.3 50.0 0.5 99.0 22.9
29 7 48.1 55.0 20.0 66.7 17.3
30 164 51.0 50.0 0.7 97.9 19.6
31 185 50.7 50.0 0.5 99.0 20.8
32 185 49.0 50.0 4.9 99.0 19.5
33 125 50.7 50.0 0.2 99.9 23.2
34 153 51.6 50.0 0.7 99.9 19.1
35 246 47.3 50.0 2.3 99.3 22.1
36 605 48.0 50.0 0.3 99.9 23.8
37 217 49.4 50.0 5.0 99.7 19.5
38 88 43.8 49.0 1.2 90.0 22.6

25, 39 303 53.3 50.0 0.2 99.8 22.3

 

91

Table 2.20: Intellectual Property Rights and Equity Ownership Structures

 

 

 

(Pure Joint Ventures)
I II III IV V VI
Ordinary Least Squares. Dep. Variable= log(equity)
ipr -.149"‘ * "‘ -.155”"‘ -.220" " -.227"* -.227* ** -.240* * f
(.020) (.021) (.028) (.032) (.032) (.032)
log h] .089 .307" .287“I .288” .434”*
(.111) (.128) (.134) (.138) (.149)
lg dp -.047*** -.047m -.047m -.053m
(.013) (.013) (.017) (.017)
phone .048 .047 .474"
(.100) (.121) (.204)
fdi -.000 .021
(.031) (.032)
black .125*"
(.048)
dummies for location, year, and industry.
Obser. 3.519
R2 .14 .14 .14 .14 .14 .14

 

Note: Standard errors in parenthesis.

(**, and "* are 5%, and 1% signiﬁcance levels, respectively.)

92

 

Table 2.21: Affiliates in the OECD and Equity Shares (%)

 

 

Code Observation Mean Median Min St. dev.
20 55 71.6 75.0 1.2 30.6
22 57 85.3 100.0 3.4 24.8
23 70 87.6 100.0 33.3 20.5
24 12 71.9 82.0 19.0 33.2
26 7 98.1 100.0 86.8 4.9
27 30 82.2 100.0 11.9 29.9
28 69 64.1 90.0 0.5 39.9
29 1 100.0 100.0 100.0 -
3O 26 71.9 100.0 0.7 36.3
31 34 82.8 100.0 0.4 29.5
32 20 75.0 100.0 10.2 33.3
33 35 71.8 100.0 0.2 38.6
34 42 79.8 100.0 0.7 31.4
35 115 72.3 100.0 2.4 36.5
36 320 72.3 100.0 0.5 36.9
37 49 70.1 82.3 5.2 30.9
38 60 66.0 94.9 1.2 40.1

25, 39 70 80.0 100.0 1.9 31.3

 

Total 1,074 74.5 100.0 0.2 34.5

93

Table 2.22: Afﬁliates in the U.S. and Equity Shares (%)

 

 

 

Code Observation Mean Median Min St. dev.
20 26 74.2 100.0 4.7 33.3
22 45 88.5 100.0 20.0 21.3
23 51 85.7 100.0 33.3 22.4
24 3 83.3 100.0 50.0 28.8
26 5 100.0 100.0 100.0 -
27 16 80.8 100.0 11.9 31.9
28 43 70.4 100.0 0.7 40.1
30 10 80.5 100.0 10.0 35.4
31 16 90.2 100.0 46.0 18.5
32 6 82.8 100.0 33.3 28.2
33 20 72.7 100.0 0.2 40.5
34 25 82.8 100.0 0.7 31.1
35 75 72.6 100.0 2.4 37.7
36 194 71.7 100.0 0.5 38.6
37 12 79.3 100.0 5.2 36.0
38 39 58.8 100.0 1.2 44.3

25, 39 46 79.4 100.0 1.9 31.2

Total 632 75.5 100.0 0.2 35.5

 

Table 2.23: Descriptive Statistics for Market Competition (Afﬁliates in the OECD)

 

 

Observation Mean St. dev. Min Max

log(equity) 1,074 4.01 1.07 -1.60 4.60
log(hhi) 1,072 4.55 0.75 2.14 6.68
log(k/l) 1,072 1.79 0.26 1.14 2.81
log(scale) 1,072 2.10 0.22 1.74 3.28
log(rnd / sales) 1,072 -1.89 0.39 -2.79 -l .41
log(h / I) 1,072 -O.37 0.20 -0.74 -0.02
log(vad /sales) 1,072 -0.29 0.06 -074 -0.17

 

Table 2.24: Descriptive Statistics for Market Competition (Afﬁliates in the U.S.)

 

 

Observation Mean St. dev. Min Max

log(equity) 632 3.99 1.13 -1.61 4.61
log(hhi) 632 4.51 0.65 2.14 6.68
log(k/l) 632 1.78 0.27 1.14 2.32
log(scale) 632 2.07 0.19 1.74 2.81
log(rnd/saleS) 632 -1.89 0.39 -2.79 -1.41
log(h /l) 632 -0.37 0.21 -0.74 -0.02
log(vad/sales) 632 -0.28 0.06 -042 -0.17

 

95

 

Table 2.25: Market Competition and Equity Ownership (All Affiliates)

 

I H HI IV V VI

 

Dep. Variable= log(equity)

 

log(hhi) -.021*** -013" -.013 .017 .017 .018
(.008) (.008) (.011) (.011) (.011) (.011)

1°ch /1) -.192*** -.120*** -.072* -.072* -.085**
(.020) (.029) (.040) (.040) (.042)

log(scale) -.160*** -. I 86*" -.188*** -.218***
(.046) (.049) (.049) (.057)
log(rnd / sales) -.039* -.048 -.045
(.022) (.032) (.032)
log(h / I ) l .024 .070
(.058) (.074)

log(vad / sales) -.182

(.1 84)
dummies for location, and year.
Obser. 8,91 7
R2 0.03 0.04 0.04 0.05 0.05 0.05

 

Note: Standard errors in parenthesis.

(*, **, and *** are 10%, 5%, and 1% signiﬁcance levels, respectively.)

96

Table 2.26: Market Competition and Equity Ownership (Affiliates in the OECD)

 

 

 

I 11 HI IV V VI
Dep. Variable= log(equity)
log(hhi) -.031 -.000 -.022 .038 .028 .030
(.043) (.043) (.063) (.070) (.070) (.070)
log(k /l) -.584*** -.645*** -.308 -.309 -.406
(.123) (.172) (.239) (.239) (.250)
log(scale) .131 -.114 -.091 -.301
(.264) (.290) (.291) (.334)
log(rnd / sales) -.248** -. I 31 -.095
(.122) (.161) (.163)
log(h / l) -.288 .052
(257) (371)
log(vad / sales) -1.339
(1.052)
dummy for year.
Obser. 1,072
R2 0.02 0.03 0.04 0.04 0.04 0.05

 

Note: Standard errors in parenthesis.

C”, and *** are 5%, and 1% signiﬁcance levels, respectively.)

97

Table 2.27: Market Competition and Equity Ownership (Afﬁliates in the U.S.)

 

 

I II 111 IV V VI
Dep. Variable= log(equity)

log(hhi) -.082 -.050 -039 .107 .074 .072
(.068) (.068) (.088) (.096) (.098) (.098)

log(k/l) -.640*** -.606*** .212 .150 .029
(.159) (.232) (.323) (.325) (.348)

log(scale) -.079 -.673 -542 -.720
(.395) (.424) (.432) (.469)

log(rnd / sales) -.603*** -.337 -.312
(.166) (.236) (.238)

log(h/l) . -.563 -.165
(356) (545)

log(vad / sales) -1 .452
(1.507)

Obser.
R2 0.00 0.03 0.03 0.05 0.05 0.05

 

Note: Standard errors in parenthesis.

( *** is 1% signiﬁcance level.)

98

‘1

REFERENCES

Antras, Pol (2003) “Firms, Contracts, and Trade Structure.” Quarterly Journal of
Economics, 118 (4), 1374-1418.

Asiedu, Elizabeth, and Hadi Salehi Esfahani (2001) “Ownership Structure in Foreign
Direct Investment Projects,” Review of Economics and Statistics, 83 (4), 647-662.

Blomstrdm, Magnus, and Ari Kokko (1997) “The Impact of Foreign Investment on Host
Countries: A Review of the Empirical Evidence,” World Bank Policy Research
Working Paper #1745.

 

, and Fredrik deholm (1999) “Foreign Direct Investment,
Technology Transfer and Spillovers: Does Local Participation with Multinationals
Matter?” European Economic Review, 43, 915-923.

Branstetter, Lee G, Raymond Fisman, and C. Fritz Foley (2005) “Do Stronger

Intellectual Property Rights Increase lntemational Technology Transfer?; Empirical
Evidence from U.S. Firm-Level Panel Data,” NBER Working Paper #11516.

Cohen, Wesley M., and Steven Klepper (1992), “The Anatomy of Industry R&D
Intensity Distribution,” American Economic Review, 82(4), 773-799.

Dimelis, Sophia, and Helen Louri (2002) “Foreign Ownership and Production
Efficiency: A Quantile Regression Analysis,” Oxford Economic Paper, 54, 449-469.

Ethier, Wilfred J., and James R. Markusen (1996) “Multinational Firms, Technology
Diffusion and Trade,” Journal of International Economics, 41, 1-28.

The Export-Import Bank of Korea (2003) Overseas Direct Investment Statistics
Yearbook, the Export-Import Bank of Korea.

Gatignon, Hubert, and Erin Anderson (1988) “The Multinational Corporation’s Degree
of Control over Foreign Affiliates: An Empirical Test of a Transaction Cost
Explanation,” Journal of Law, Economics, and Organization, 4(2), 305-336.

Ginarte, Juan C., and Walter G Park (1997) “Determinants of Patent Rights: A Cross-
National Study,” Research Policy #26, 283-301.

99

T

Gomes-Casseres, Benjamin (1989) “Ownership Structures of Foreign Affiliates: Theory
and Evidence,” Journal of Economic Behavior and Organization, 11 (1), 1-25.

 

(1990) “Firm Ownership Preferences and Host Government
Restrictions: An Integrated Approach,” Journal of International Business Studies,
21 (1 ), 1-22.

Hall, Robert E. and Charles 1. Jones (1999) “Why Do Some Countries Produce So
Much More Output per Worker than Others?” Quarterly Journal of Economics,
114(1), 83-116.

Head, Keith, and Thierry Mayer (2004) “Market Potential and the Location of Japanese

Investment in the European Union,” The Review of Economics and Statistics, 86
(4), 959-972.

Kim, June-Dong, and In-Soo Kang (1997) “Outward FDI and Exports: The Case of
South Korea and Japan,” Journal of Asian Economics, 8 (1), 39-50.

Kobrin, Stephen J. (1987) “Testing the Bargaining Hypothesis in the Manufacturing
Sector in Developing Countries,” International Organization, 41 (4), 609-63 8.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny
(1998) “Law and Finance,” The Journal of Political Economy, 106(6), 1113-1155.

Lee, Jeong—Yeon, and Edwin Mansﬁeld (1996) “Intellectual Property Protection and
U.S. Foreign Direct Investment,” Review of Economics and Statistics, 78 (2), 181-
186.

Mansﬁeld, Edwin (1994) “Intellectual Property Protection, Foreign Direct Investment,
and Technology Transfer,” World Bank, IFC Discussion Paper #19.

 

(1995) “Intellectual Property Protection, Direct Investment, and
Technology Transfer: Germany, Japan, and the United States,” World Bank, IFC
Discussion Paper #27.

 

* , and Anthony Romeo (1980) “Technology Transfer to Overseas
Affiliates by U.S.-Based Firms,” Quarterly Journal of Economics, 95 (4), 737-750.

Markusen, James R. (1995) “The Boundaries of Multinational Enterprises and the
100

Theory of International Trade,” Journal of Economic Perspectives, 9 (2), 169-189.

 

(2001) “Contracts, Intellectual Property Rights, and Multinational
Investment in Developing Countries,” Journal of International Economics, 53,
189-204.

Miles, Marc A., Edwin J. Feulner, Jr., Mary Anastasia O’Grady, and Ana 1. Eiras (2003)
Index of Economic Freedom, Heritage Foundation.

McCalman, Phillip (2004) “Foreign Direct Investment and Intellectual Property Rights:
Evidence from Hollywood’s Global Distribution of Movies and Videos,” Journal
of International Economics, 62, 107-123.

Neary, J. Peter (2002) “Foreign Direct Investment and the Single Market,” The
Manchester School, 70 (3), 291 -3l4.

Nicholson, Michael W. (2002) “Intellectual Property Rights, Intemalization and
Technology Transfer,” Federal Trade Commission, Working Paper.

Smarzynska, Beata K. (2000) “Technological Leadership and Foreign Investors’ Choice
of Entry Mode,” World Bank Policy Research Working Paper #2314.

 

(2004) “The Composition of Foreign Direct Investment and
Protection of Intellectual Property Rights: Evidence from Transition Economies,”
European Economic Review, 48 (1), 39-62.

UNCTAD (2002) World Investment Report, United Nations.

U.S. Department of Commerce (2001) “Concentration Ratios in Manufacturing: 1997
Economic Census,” Department of Commerce.

Vishwasrao, Sharmila (1994) “Intellectual Property Rights and the Mode of Technology
Transfer,” Journal of Development Economics, 44, 381 -402.

Wooldridge, Jeffrey M. (2000) Introductory Econometrics - A Modern Approach, South-
Westem College Publishing.

101

CHAPTER III

Licensing, Foreign Direct Investment,

and Government Policies

1. Introduction

Ronald Coase (1937) raised one basic question in economics, “Why are there firms
even though production can be organized through the price mechanism?” He explained
that “the main reason why it is proﬁtable to establish a f'um would seem to be that there
is a cost of using the price mechanism (p. 390).” Although all activities in the market
can be organized through the price mechanism, the costs for contracts are considerable
when the activities are complex.

He said that contracts are greatly reduced if there is a ﬁrm. Then, what then
determines the size of the ﬁrm? He emphasized that, at the margin, the costs of
organizing within the ﬁrm will be equal either to the costs of organizing in another ﬁrm
(plant) or to the costs involved in leaving the transaction to be “organized” by the price
mechanism, at the margin (p. 404).

Based on the principle of marginalism, a multinational ﬁrm in the source (or
developed) country needs to choose between licensing and foreign direct investment1 as
its entry mode to the host country. However, the host (or developing) country would
prefer licensing to foreign direct investment in the early stage of development. For

example, Japanese government agencies restricted inﬂows of foreign direct investment,

‘

I If a multinational ﬁrm chooses licensing (an arm’s length arrangement), then it transacts business with
local company in the host country outside its boundaries. However, if the ﬁrm chooses foreign direct
Investment (internalization), then it transacts business within its boundaries.

102

‘1‘"

but they aggressively encouraged licensing until 1970. 2 Similarly, the Korean
government generally discouraged foreign direct investment, but it was rather generous
toward licensing during the 19705 and 805 (Pack and Saggi, 1997).3

Each country has its own standards of intellectual property rights protection (Ginarte
and Park, 1997).4 The standards of intellectual property rights protection would be
positively related to the enforceability of licensing contracts and, thus, licensor’s rent
share.5 6 Although many papers have analyzed the effects of intellectual property rights
protection on technology spillovers,7 few papers have studied the effects of contract
enforceability,8 plant-level ﬁxed costs, and market structures in the host country on the
ﬁrm’s choice between licensing and foreign direct investment.

The ﬁrm’s choice of entry mode also inﬂuences the social welfare of the host
country. Hence, the host country tends to affect a multinational ﬁrm’s choice of entry

mode by changing the degree of intellectual property rights protection. In addition,

 

2 Before 1970, Japan was underdeveloped compared to other OECD countries such as the United States
and the United Kingdom.

3 Vishwasrao (1994) also emphasized that Southern countries have typically imposed numerous
restrictions on foreign direct investment, ranging from nationalization to compulsory licensing and
restrictions on repatriation of proﬁts (p. 383). For example, India, Brazil, Argentina, and Japan, have all
had policies which discouraged foreign direct investment in favor of licensing.

’ To measure the standards of intellectual property rights protection of 110 countries, Ginarte and Park
(1997) took into account ﬁve categories of patent laws: (1) extent of coverage, (2) membership in
international patent agreement, (3) provisions for loss of protection, (4) enforcement mechanisms, and (5)
duration of protection.

5 We consider output royalty, rent sharing between a multinational ﬁrm and licensee. Most licensing
contracts include an output royalty (Gallini and Wright, 1990). But we do not consider various types of
licensee’s payment such as a ﬁxed fee and nonlinear royalty schemes for simplicity.

° The literature emphasized that the licensor can capture only a fraction of the gains from licensing
agreement. For example, Chin and Grossman (1990), Contractor (1985), Horstrnann and Markusen
(1987), Yang and Maskus (2001).

7 For example. Markusen (1995) showed that low standards of intellectual property rights are likely to be
associated with foreign direct investment rather than licensing being chosen as the governance structure
due to technology spillovers. Empirically, Mansfield and Romeo (1980) showed that multinational ﬁrms
tend to transfer their newest technologies overseas through affiliates rather than licensing since ﬁrms fear
technology spillovers. They conﬁrmed that the mean age of the technologies transferred to overseas
affiliates in developed countries was signiﬁcantly less than the mean age of the technologies transferred
to overseas affiliates in developing countries. This is because many newer technologies are inappropriate
for developing countries, or are difficult and expensive to transfer there.

8 What the stronger enforceability of licensing contracts the host country has might mean that the licensor
can execute the licensing arrangement more efficiently. On the contrary, if contracts are poorly enforced,
then the licensee can abrogate the licensing agreement partially (or wholly) and obtain a part (or the
Whole) of the licensor’s rents. We assume that the variable for the host country’s contract enforceability is
cOntinuous.

103

many countries have used various policies to promote inward foreign direct investment,
such as directly subsidizing multinational ﬁrms and/or extending tax holidays (Hanson,
2001). We give the economic rationale for those various host countries’ policies to
increase social welfare.

We develop a simple model to study a multinational ﬁrm’s choice between licensing
and foreign direct investment. Multinational ﬁrms react optimally to various policies
and market conditions they face in host countries. We examine the amount of
technology transfer to the host country, welfare implications of each entry mode, and
the country’s strategic policies that cause the multinational ﬁrm to choose a desirable9
entry mode.

The likelihood that local companies in the host country absorb the multinational
ﬁrm’s advanced technology depends on the nature of technology and the characteristics
of the country in which they are located. This paper shows that the actual policy
instrument used by the host country is dependent on the capability to absorb the
multinational ﬁrm’s technology. One interesting result is that along with other factors
such as contract enforceability, the market structures in the host country affect a
multinational ﬁrm’s decision for entry mode and, hence, the country’s welfare.lo

We give a novel explanation instead of technology spillovers with regard to why
developing countries encouraged licensing rather than inward foreign direct investment.
Contrary to Hanson (2001), we show that the host country’s direct subsidies to induce a
multinational ﬁrm to choose foreign direct investment can beneﬁt both the host country

and the multinational ﬁrm under certain conditions. In addition, we demonstrate that the

 

9 A desirable entry mode is deﬁned as one of entry modes between licensing and foreign direct
investment that gives greater welfare to the host country than the other entry mode.

1° Dasgupta (1986) pointed out that market structures should not be taken as given in a theory related to a
ﬁrm’s technology. Thus, a multinational ﬁrm can affect the market structures of the host country by
changing the amount of technology transfer. Zigié (1998) also focused on the market structures. In
contrast to aforementioned papers, we endogenize the multinational ﬁrm’s entry mode and the extent of
technology transfer given market structures in the host country.

104

f" .

host country never chooses perfect contract enforceability when one incumbent
company already exists in the host country and can compete with the multinational ﬁrm.

Markusen (1995) showed that multinational ﬁrms are important in industries that
demonstrate high levels of R&D relative to sales and have a large share of professional
and technical workers in their workforces. Hence, multinational ﬁrms can increase their
competitiveness in the host country by transferring a greater amount of technology to
their affiliates. However, technology transfer would incur non-negligible costs for the
multinational ﬁrms. Teece (1977) emphasized that the costs of the technology transfer,
which are deﬁned to include both transmission and absorption costs,ll might be
considerable when the technology is complex and the recipient does not have the
capability to absorb it. '2

We consider only greenﬁeld foreign direct investment rather "than acquisition of
local company.13 Though the degree of plant-level ﬁxed costs may play a role in
determining the ﬁrm’s choice between licensing and foreign direct investment, other
considerations are important. If a multinational ﬁrm chooses licensing contracts with an
incumbent local company instead of foreign direct investment, then it has to share rents
with its local partner because of imperfect contract enforceability. However, if the firm
decides to enter the host country by foreign direct investment, then it pays the plant-
level ﬁxed costs and increases the degree of competition if any incumbent local
companies make the same variety of the multinational ﬁrm.

Some related literature focuses on the ﬁrm’s choice to enter the host country and
affect the host country’s welfare. Markusen (2001) who considered the ﬁrrn’s choice

between exporting and a subsidiary found that stricter contract enforcement is better for

 

1' The advisory and consulting costs are also included since these activities are necessary if the
technology is to be adjusted to the local conditions and requirements.

‘2 Teece found that on average costs of technology transfer were 19 percent of total project costs, in some
cases amounting to as much as 59 percent.

‘3 A greenﬁeld is deﬁned as setting up a new affiliate in the host country, and an acquisition is the
takeover of an incumbent company in the host country (Vermeulen and Barkema, 2001).

105

 

the multinational ﬁrm. He also showed that both multinational ﬁrm’s proﬁts and the
host country’s welfare are improved by the institution of contract enforcement if stricter
contract enforceability leads to a mode shift from exporting to production within the
host country. However, he did not consider the effects of market structures on the ﬁrm’s
choice and the host country’s welfare, nor did he consider the extent of technology
transfer. Horstrnann and Markusen (1992) showed that market structures can be
endogenized and that the host country’s welfare is dependent on its market structure.
Chin and Grossman (1990) studied the incentive that a government in the South has
to protect the intellectual rights of Northern ﬁrms. They examined the consequences of
the decision taken by the South for welfare in the North and for efficiency of the world
equilibrium with the North-South trade model. They considered two extreme cases,
complete intellectual property rights protection versus complete violation. Zigié (1998)
extended and complemented their model of intellectual property rights. He introduced
the “R&D with spillovers” type of model and reconsidered the welfare effects of
intellectual pr0perty rights protection in the North-South trade relationship.M Chin and
Grossman, and Zigié focused on the strength of intellectual property rights protection
and its welfare outcomes of the North, the South, and the whole world. However,
neither paper considered that the Southern government can set policies strategically.
Petit and Sanna-Randaccio (2000) examined the impact of the ﬁrm’s mode of
foreign expansion on the incentive to innovate as well as the effects of R&D activities
and technological spillovers on the ﬁrms’ international strategy. They showed that the
spillover parameter can be an important determinant of ﬁrm’s international strategies.
They considered a model with two identical countries and two ﬁrms, and then compared

R&D investments that resulted from various market structures.

 

'4 He analyzed the model under assumptions that R&D spillovers are deﬁned as the leakage of important
pieces of technical information which can be used by the recipient at zero marginal costs. The intensity of
spillovers is interpreted to reﬂect the strength of intellectual property rights protection.

106

Horstrnann and Markusen (1987) demonstrated that a multinational ﬁrm might
choose a strategy of direct investment rather than licensing since the ﬁrm cannot
monitor a licensee to guarantee that the licensee maintains the reputation. They
emphasized the multinational ﬁrm’s concerns on contracts. Vishwasrao (1994) showed
that the lack of adequate intellectual property protection in the South may encourage
foreign direct investment instead of licensing. Saggi (1999) built a two-period duopoly
model to study a multinational ﬁrm’s choice between licensing and foreign direct
investment and the impacts of these modes on the incentives for innovation of that ﬁrm
and its local rival. He also analyzed the host country’s welfare and derived policy
implications. However, Horstrnann and Makusen, Vishwasrao, and, Saggi did not
examine the effects of the host country’s market structures and the country’s strategic
policies.

The remainder is organized as follows. Section 2 describes the features of a basic
model with two countries: (1) a source country with a multinational ﬁrm and (2) a host
country. The host country is segmented from the rest of the world. Sections 3 to 5
consider three separate cases: one with no incumbent local company in the host country
that can compete with the multinational ﬁrm, one with one competent local company in
the host country, and one with two competent local companies in the host country. The

last section concludes.

2. The Basic Model

We assume that there are two countries, a source and a host: a multinational ﬁrm of
variety i is located in the source country. We assume that the host country is separated
from the rest of the world because of distance and high tariffs. Hence, we consider the

multinational ﬁrm’s choice between foreign direct investment and licensing to enter the

107

host country’s market because exporting is an inferior choice for the ﬁrm to foreign
direct investment and licensing.

The multinational ﬁrm in the source country, assumed to have advanced technology,
must decide between licensing and foreign direct investment. We consider only
greenﬁeld foreign direct investment rather than acquisition of a local company although
the host country has incumbent local companies that can compete with the ﬁrm. This
restriction15 could be justiﬁed either because the host country prevents foreign investors
from acquiring the local companies or because the capital market in the host country is
not developed enough for foreign investors to sufficiently acquire an equity share of a
local company to control that company.

The inverse market demand for variety i in the host country is assumed to be

linear
P=A—Q (30

where A is a measure of market size and Q is total output in the market. When the
multinational ﬁrm enters the host country by either licensing contracts or foreign direct
investment, the ﬁrm can produce the variety 1' by initial marginal cost a. That is, a
is the marginal cost when there is no technology transfer from the multinational ﬁrm. '6
However, if the ﬁrm transfers its advanced technology to the host country, then the
marginal cost for the variety i is reduced to a -t , where t is the extent of
technology transfer. Following Teece (1977), we assume that the multinational ﬁrm

must pay the costs of technology transfer‘7

C(t) =%yt2, 7 >0 (3.2)

 

’5 In addition, a local company might be reluctant to be acquired by any multinational ﬁrms because of
concerns of local company’s director over losing control rights.

'6 It is assumed that A > a .

'7 Norback (2001) used similar quadratic technology transfer cost function. Petit and Sanna-Randaccio
(2000) assumed the quadratic form of R&D expenditure.

108

where 7 denotes the efficiency of technology transfer. This 7 is positively related to
the complexity of transferred technology, but negatively related to the host country’s
capability to absorb the transferred technology. The country’s capability to absorb the
transferred technology can be increased by government policy, such as subsidies for
higher education or for local companies’ investments for research and development
(R&D), targeted technology inducements, and competition policies.18 We assume that
the host country cannot improve the absorptive capacity (7) in the short-run.

Consider respective beneﬁts and costs of two entry modes, licensing and foreign
direct investment. If the multinational ﬁrm decides to enter the host country by an arm’s
length arrangement, then the ﬁrm can avoid incurring plant-level ﬁxed costs, F , to set
up a new plant as an affiliate. There might be ﬁxed costs under licensing contracts. Thus,
plant-level ﬁxed costs, F , might be interpreted to represent the difference between the
costs under foreign direct investment and those under licensing since establishing an
affiliate may incur more ﬁxed costs.

We assume that the important cost for licensing is that the multinational ﬁrm cannot
obtain whole operating proﬁts19 due to imperfect contract enforceability. Yang and
Maskus (2001) assumed that stronger intellectual property rights protection would
increase the licensor’s rent share. Gallini and Wright (1990) demonstrated that when
imitation is possible and there is asymmetric information, the licensor would share the
rents with its licensee. We assume that the host country can change the enforceability of
licensing contracts and thus the share, 19, of operating proﬁts for the multinational ﬁrm

(where 66 (0,1]).20 In other words, the multinational ﬁrm is assumed to obtain 0

 

" Pack and Saggi (1997), and Maskus et a1 (2004).

'9 See Contractor (1985), and Chin and Grossman (1990), Horstrnann and Markusen (1987). For example,
Horstrnann and Markusen emphasized that the licensing agreement must provide incentives for the
licensee and, thus, transfer some of returns to the licensee.

2° If 9 = 0 , then the multinational ﬁrm never chooses licensing.

109

r.

share of operating proﬁts”1 under licensing.

The parameter 6 that represents the contract enforceability is a host country’s
policy instrument.22 By amending related laws and provisions, the host country can
change the level of contract enforceability and, hence, the share of operating proﬁts for
the multinational ﬁrm. If local companies in the host country know ex ante that a future
licensee can obtain (1—6) share of operating proﬁts when the multinational ﬁrm
chooses one of the local companies as a licensing partner, then the multinational ﬁrm
can capture all operating proﬁts by auctioning the (1-6) share of operating proﬁts to
the highest bidder.

However, that is impossible because we assume asymmetric information between
the multinational ﬁrm and local companies as Gallini and Wright (1990) did. They
assumed asymmetric information that a licensor had superior pre-contractual
information regarding the economic value of technology, more than its licensee. Then,
they showed that stronger intellectual property rights were associated with larger
licensor’s rent share.

Similarly, we assume that only the multinational ﬁrm of variety i knows the
complexity of technology. That is to say, the ﬁrm is much more informed about its own
technology than local companies of variety j (where i: j) in the host country.23
Thus, it is impossible that the multinational ﬁrm auctions the licensee’s share of
operating proﬁts to local companies; and then the ﬁrm obtains all operating proﬁts

under the choice of licensing.

 

2' As we can see in the last part of this section, the multinational ﬁrm transfers its technology to its
licensee in the third stage. In the fourth stage, the licensee produces the variety i by using the
transferred technology and makes operating proﬁts. Thus, the technology transfer costs become sunk
costs in the fourth stage, and the licensee does not need to consider the sunk costs in the fourth stage since
the multinational ﬁrm bears all the sunk costs.

22 Markusen (2001) assumed that the multinational ﬁrm / licensee must pay a penalty if the ﬁrm / licensee
defects their contracts, and he regarded the penalty as a measure of contract enforcement (as a policy
instrument).

23 For the cases in which one or two local companies in the host country produce the variety i as in the
sections 4 and 5, we extend the assumption of asymmetric information in the following section.

110

 

By taking into account of the complexity of the technology and host country’s
contract enforceability, only the multinational ﬁrm of variety i knows its share, 6, of
operating proﬁts ex ante. In section 3, in which no local company of variety 1' exists in
the host country, the multinational ﬁrm cannot auction the licensee’s share of operating
proﬁts to the local companies of variety j since the local companies do not know ex
ante that a licensee can acquire (1—19) share of operating proﬁts.

Contrary to licensing, if the multinational ﬁrm chooses foreign direct investment,
the ﬁrm does not need to concern itself with the host country’s contract enforceability.
As we explained, however, the ﬁrm must pay plant-level ﬁxed costs, F . In addition, if
any competitive local companies of variety i are in the host country, greenﬁeld foreign
direct investment increases the competition in the market. In summary, the multinational
ﬁrm must consider respective beneﬁts and costs of two entry modes when it decides an
entry mode: 6 share of operating proﬁts under an arm’s length arrangement, ﬁxed
costs under foreign direct investment, and market structures in the host country.

The host country can use various policies to induce the multinational ﬁrm to choose
a desirable entry mode, such as changing the share of operating proﬁts for the
multinational ﬁrm (6) by altering contract enforceability,” direct subsidies, and direct
restrictions on any entry mode. For example, if the host country decides to encourage
inward foreign direct investment, then the country can offer direct subsidies for the
multinational ﬁrm.25 However, we assume that the host country prefers changing

contract enforceability to giving direct subsidies to the multinational ﬁrm as a policy

 

2’ A change in contract enforceability affects both the multinational ﬁrm’s choice and the local
companies’ behavior in the host country. In other words, since the effects of the host country’s change in
contract enforceability are very extensive, the host country would consider the effects on the local
companies as well as on the multinational ﬁrm. However, since the objective of this chapter is to give an
economic rationale for the difference in contract enforceability among the host countries, we focus on the
effects of the change in enforceability on the multinational ﬁrm.

25 By using such as partial (or complete) exemptions from corporate taxes, the host country can subsidize
the multinational ﬁrm to choose foreign direct investment rather than licensing.

lll

 

instrument. This is because by changing enforceability26 instead of subsidizing, the host
country can avoid the ﬁnancial burden incurred by direct subsidies. Whether the host
country encourages licensing or foreign direct investment depends on relative extents of
the country’s welfare between two entry modes.

The game in this model consists of four stages. In the ﬁrst stage, the host country
sets up policies to promote licensing or foreign direct investment. The host country can
choose both the level of contract enforceability and direct subsidies. In the second stage,
the multinational ﬁrm decides between licensing and foreign direct investment as its
entry mode into the host country. In the third stage, given the entry mode, the
multinational ﬁrm transfers its technology to the host country. In the ﬁnal stage, the

variety i is produced in the host country.

3. Case 1: Without Competent Local Company

As a benchmark, we begin with the case in which no local company is producing
variety i in the host country. This implies that the multinational ﬁrm can be a
monopolist in the host country regardless of its choice of entry mode between licensing
and foreign direct investment. As a rule, we study this case by backwards induction. If
the multinational ﬁrm makes licensing contracts with a new local company of variety

j , then the proﬁt27 for the ﬁrm is given by
1
tr,“ =6((A—q,)q,-(a—t,)q,)—5n,’ (3.3)

where subscript l and superscript m represent licensing and the multinational ﬁrm,

respectively.

 

2‘ To be strict, there must be other costs to change 6. For example, the host country must amend related
laws to change 6. For simplicity, however, we assume that the host country prefers changing 6 to
giving direct subsidies. But we will examine the case where the host country prefers direct subsidies to
changing contract enforceability.

27 We assume that all proﬁts are non-negative in this chapter.

112

 

On the contrary, if the firm decides to enter the host country through foreign direct

investment, then its proﬁt28 reads as
Irm=(A— ) -(a—t ) —lyt2-F (34)
f qf qf f qf 2 f '

where subscript f represents foreign direct investment. In the last stage, the ﬁrm acts
to maximize its proﬁt by choosing output level. Simple derivation gives the ﬁrm’s

output of each mode as following.

i=1,
2 f

Q2

When inserting above formula into equations (3.3) and (3.4), we represent proﬁt

functions by parameters and technology transfer (t, , i =1, f ).

In the third stage, the ﬁrm chooses the extent of technology transferz‘9

t1 =———6(A‘a) (3.5)
27—6
A —a
t = —— 3.6
f 27 _1 ( )
By using the above, we obtain ﬁrm ’5 outputs of two entry modes
7(A -a)
q =—-—— (3.7)
I (Zr-9)
7(A -a)
= 3.8
f 27 _1 ( )

From equations (3.5) and (3.6), we show that the ﬁrm transfers less extent of

technology under licensing than under foreign direct investment.30 The intuition of this

 

2' When the host country decides to give direct subsidies for the multinational ﬁrm to choose foreign
direct investment in the ﬁrst stage, then the multinational ﬁrm’s proﬁt becomes a? = (A—q [”7 f

-(a - t f )q f — ‘2L rt; — F + S , where S represents host country’s direct subsidies.

29 For the second order conditions to be satisﬁed. assume that 27 — 6 > 0 , and 27 —1 > 0 .
3° If 6 = 1 , the extents of transferred technology are identical under both entry modes. For simplicity, if

113

r .

result is simple and straightforward. If the ﬁrm chooses an arm’s length arrangement,
then it must share the operating proﬁts with its licensee due to imperfect contract

enforceability even with paying all technology transfer costs. This leads to a higher

marginal cost under licensing (a—tl) relative to that under foreign direct investment

( a —t ). Accordingly, the ﬁrm produces less output under licensing than under foreign
f

direct investment. In addition, a simple comparative static analysis conﬁrms several
intuitive results. The ﬁrm transfers more technology to the host country under an arm’s
length arrangement, as the contract enforceability becomes stricter.“ Also, the ﬁrm
produces more output under licensing as the host country strengthens contract
enforceability. 3’2

In the second stage, whether the ﬁrm chooses licensing or foreign direct investment
depends on relative proﬁts between licensing and foreign direct investment. The ﬁrm’s
proﬁts under licensing and foreign direct investment are represented by parameters and

the host country’s contract enforceability (6).

7H" 2M (3 9)
1 2(27—9) '
”m =M_F (310)

f 2(27—1) '

From equation (3.9), we can show that the multinational ﬁrm’s proﬁt under

 

we do not mention the range of 6 speciﬁcally, then we usually consider the case in which 0 6 (0,1)
when comparing the results between licensing and foreign direct investment.

3‘ az,/ao=2y(A-a)/(0-2y)2 >0.
’2 aq,/aa=y(A-a)/(2y-9)2 >0.

”4

 

licensing is increasing with respect to contract enforceability.” Thus, strict contract
enforceability attracts an arm’s length arrangement rather than foreign direct investment.

In addition, a rise in the host country’s capability to absorb the multinational ﬁrm’s

advanced technology (lower 7) increases the ﬁrm’s proﬁts irrespective of entry modes,

an’" _ 2 2 67:“ _ _ 2
I =-(A a) 2 <0 and f = (A a)2<0.
67 2(27 -9) 67 2(27 —1)

 

since

 

Equations (3.9) and (3.10) give a critical value of plant-level ﬁxed costs, F_‘ ,34

17.; 7204-00204)
(Zr-IXZr-B)

 

(3.11)

such that for F > F , the multinational ﬁrm chooses licensing instead of foreign direct

investment.35 This critical value of ﬁxed costs is decreasing with respect to 6.

6F -72(A—a)2
— = ———2— < 0
‘39 (Zr-9)
This is intuitive because if contract enforceability in the host country increases, an
arm’s length arrangement becomes more proﬁtable to the multinational ﬁrm than
foreign direct investment. We also ﬁnd that an improvement in the host country’s

capability to absorb advanced technology does not decrease the critical value of ﬁxed

 

a::,’"_72(A-a)2
50 (6-27)2

3‘ The second order conditions (27 — 6 > 0 , and 27 —l > 0) guarantee that F is positive.

 

’3 This can be veriﬁed by > O.

35 Similarly, we can ﬁnd a critical value of contract enforceability, 6 , such that for 6 >6 , the
multinational ﬁrm chooses licensing rather than foreign direct investment.

2
_ A— -2F2 -1 _
7; a) 2 ( 7 ) .Thus, we can show that the effects of 7 on 6 are negative.
7 (A-a) —F(27-1)

 

a5 2F(27-l)(F(27-l)-7(A-a)2)

 

= < 0 (the nominator is negative since
a 2 2 2
7 (r (A—a) -F(27-l))
2
A _
”7 = _7_(_a)__ F > 0 ).
2(27 —1)

115

costs. That is, an increase in capability attracts the multinational ﬁrm to choose foreign

direct investment when real ﬁxed costs are in the neighborhood of the critical value.

 

gr; 27(A—a)2(1—6)(0—7-76) so
67 (Zr-1)2(27-0)2

Similar to Horstrnann and Markusen (1987), we can conﬁrm that licensing is more

likely to be observed in small markets and foreign direct investment in large ones.

a_1?__ 272(A-a)(1-6) >0
6A (27—1)(27-6) —

 

Figure 3.1: The Multinational Firm’s Proﬁts under Two Entry Modes

 

 

 

 

 

 

 

 

V

 

 

 

Figure 3.1 shows the multinational ﬁrm’s proﬁts under two entry modes in {6,7r}

space. The curve for ﬁrm’s proﬁt under licensing (film ) is positive and convex,36 and

 

 

 

572m 2ﬂm
3‘ I > 0 , and 5 > 0.
66 66

116

m

2
72'! is the greatest [M EM

when 6:]. The proﬁt under foreign direct
2(27-1)

investment (It?) is described by horizontal line because it is not affected by contract

enforceability (6). In addition, 6 is a critical value of contract enforceability. The

difference between the greatest proﬁt under licensing (M ) and that under foreign direct

investment (71’?) is the plant-level ﬁxed costs (F ).37 The intuition is as follows.

Suppose that the host country sets the contract enforceability (6) stricter than the
critical level of contract enforceability (6 ). This causes the multinational ﬁrm to
choose licensing. Conversely, an increase in direct subsidies raises the possibility for
multinational ﬁrm to choose foreign direct investment over licensing.

In the ﬁrst stage, the host country sets up policies to induce the ﬁrm to choose a

desirable entry mode that gives greater social welfare. The social welfare (Wt where
i=1, f ) of the host country is deﬁned as the sum of consumer surplus (Si where

i =1, f ) and producer surplus (local company’s proﬁt38 zrlc, where superscript c

denotes the local company). In the case of a linear demand function, the consumer

Q2

surplus is expressed by —2— (where Q represents total output in the market).

The host country can set up contract enforceability, 6, and / or subsidize a part of
(or all) ﬁxed costs for the multinational ﬁrm. Before setting up any policies, the host
country compares the welfare under licensing with that under foreign direct investment.
Consumers are worse off under licensing than under foreign direct investment since the

ﬁrm transfers less extent of technology to the host country and, hence, total output is

 

’7 If the host country decided to subsidize the multinational firm to encourage foreign direct investment.
then the difference is the plant-level ﬁxed costs minus the country’s direct subsidies for the ﬁrm (F — S ).

3' In this section in which no competent local company of variety 1' exists in the host country, the host
country do not have producer surplus if the multinational ﬁrm chooses foreign direct investment.

”7

 

less under licensing than under foreign direct investment. However, under licensing, the
local company (licensee) which is the partner of the multinational ﬁrm can make proﬁt
due to imperfect contract enforceability.

The expression for social welfare under licensing is represented by

2
q
W, =(1-ﬂ)%+ﬂrrf =(1—ﬂ)—5’—+ﬂ(1—6)(A—q,—a+t,)q,
where (1— ,6) and ,6 denote the weights of consumer surplus and producer surplus in

the host country’s social welfare, respectively. If we assume that the host country set

identical weights for two surpluses ( ,6 zé), then the respective expressions of social

welfares under licensing and foreign direct investment become

= (3 -26)72(A —a)2
2(27-19)2

 

W

1 (3.12)

2 2 2
W =22_=?L=Z__(’_4‘_a)_ (3.13)
f 2 2 2(27—1)2

An increase in efﬁciency of technology transfer raises the country’s social welfare.

6W _ _ 2 6W _ _ 2
,zmze 3)(A a) (mm f: 7(A 0:) <0.

67 (27—6)3 67 (Zr-n3

 

 

However, under an arm’s length arrangement, a decrease in contract enforceability to
obtain larger share of operating proﬁts does not always increase the host country’s

welfare.

6W: = 72(A—a)’(3—2r—6)
66 (27—49)3

 

 

That is, the effects of increase in 6 on the host country’s social welfare depends on

118

 

the sign of 3—27—6. If the multinational ﬁrm can transfer its advanced technology

efficiently39 (7 <3—3Q), then an increase in contract enforceability increases the host

6W
country’s welfare [—6_6L > O]. In other words, although stricter contract enforceability

lowers the local company’s share of operating proﬁts, the gain in consumer surplus
outweighs the loss of producer surplus since stricter contract enforceability and higher
efﬁciency of technology transfer encourage the multinational ﬁrm to transfer a greater

extent of technology.

Deﬁne DsWI—W f , where D is the difference in social welfares between

licensing and foreign direct investment. If D is positive, then social welfare under
licensing is greater than under foreign direct investment. After subtracting equation

(3.13) from equation (3.12) and rearranging the result, we can simplify D

(1 —0)(9+ 872 —127 + 3)72(A —a)2
W! ’W E D = 2 2
2(27-9) (27-1)

 

(3.14)

From equation (3.14), we can ﬁnd threshold contract enforceability, 6 , such that

social welfare under licensing is identical to that under foreign direct investment when

6 = 6. If contract enforceability is perfect (6 =1), then the host country is indifferent

( D = O ) between licensing and foreign direct investment. In addition, if
6 = —872 +127—3 , the host country’s social welfare under licensing is identical to that
under foreign direct investment (D = 0 ). Since it is assumed that 6 e (O, l] and 7 > 1 ,

we can illustrate the host country’s preference for the multinational ﬁrm ’5 entry mode in

{ 7,6} space.

 

39 This can be interpreted like this, the host country can adopt the multinational ﬁrm's technology easily.

“9

In Figure 2, the segmenting curve, which divides the whole region into two regions,

B and C, is derived by the formula, 6=—872+127—3. Figure 2 shows that if

parameters are located in region B , the left side of the curve, then the host country
prefers foreign direct investment to licensing (D < 0). In region B , the multinational
ﬁrm can transfer its technology to its afﬁliate in the host country efficiently (low value
of 7). Under low value of 7, the gain in consumer surplus caused by a greater extent
of transferred technology under foreign direct investment outweighs the loss of
producer surplus that could be made under licensing. Licensing causes the multinational
ﬁrm to transfer less technology since the firm must share the operating proﬁts with the
licensee. Licensing would hurt consumers more in the host country as the technology
can be transferred more efﬁciently.

However, if the multinational ﬁrm cannot transfer its technology efﬁciently (region
C in Figure 3.2, the right side of the segmenting curve), then licensing is better entry
mode for the country (D > 0). Figure 3.2 shows that if the value of 7 is greater than
1.18, then the country always prefers licensing to foreign direct investment irrespective
of the values of 6. If the parameters are located in region C, then an arm’s length
arrangement gives more social welfare to the host country than foreign direct
investment. In other words, the licensee’s proﬁt outweighs the loss of consumer surplus.

We can also divide the region C into two sub-regions, d and e, by using the

6W

formula 3—27-6 =0 (dividing line), derived from 7361—:0' In sub-region d, the

area between the right side of the segmenting curve and the left side of the dividing line,
a rise in contract enforceability increases the host country’s social welfare under

6W

licensing [736—1 > 0]. Conversely, in sub-region e , the right side of the line, a rise in

contract enforceability decreases the host country’s social welfare under licensing

120

 

    
  
  

' 4
Licensin
(C g

CD
A
.<
v

I
E
-J""

___I)ividing Line (3 — 27 - 0 = 0)
am /69 < 0

(e)
l

:6W,/66>0
(d)

 

 

 

:0"
co
;

0.5 y 4

This model delivers a novel explanation about why developing countries
encouraged licensing rather than inward foreign direct investment in the early stage of
development. To the best of our knowledge, the literature on this issue has focused on
technology spillovers: the host country considers the difference in the degrees of
technology spillovers to local companies between two modes. However, as Blomstrbm
and Kokko (1997) asserted in their survey paper, if local companies in the developing
countries lack absorptive capacity for advanced technologies, developing economies
enjoy few positive effects of inﬂows of advanced technologies.

It is difﬁcult to conﬁrm whether the developing countries had potentials enough to
absorb advanced technologies in the early stage of development. Thus, there may have
been other reasons rather than technology spillovers for the developing countries to

have preferred licensing to foreign direct investment. Probably, there were no local

121

companies in the developing countries that could compete with multinational ﬁrms in
the early stage of development. Both less development of infrastructures and lack of
absorptive capacity could incur high technology transfer costs (region C in Figure 2).
Hence, the developing countries were likely to prefer licensing to foreign direct
investment when their economies were in the early stage of development.

As explained, the host country can use a few policy instruments, such as contract
enforceability and direct subsidies, to induce the multinational ﬁrm to choose a
desirable entry mode. Equation (3.11) shows that the host country can change the
critical value of ﬁxed costs, I“: , by altering contract enforceability, 6. We can
calculate the possible range of F- depending on the change in 6. Under perfect
contract enforceability (6 =1), the critical value of ﬁxed costs is 0. On the contrary, the

largest critical value (F m) is obtained when 6 = 0. Thus, the range of F‘ is given by

OSF<MEFM (3.15)
2(27-1)

In Figure 2, if 7 is greater than 1.18, then the host country prefers licensing to
foreign direct investment. Most developing countries might not have any local
companies that can compete with multinational ﬁrms of the developed countries and the
host country might lack enough human capital to easily absorb multinational ﬁrms’
advanced technologies. Thus, the social welfares of developing countries would be
better under licensing than foreign direct investment (region C in Figure 3.2).

To induce multinational ﬁrms to choose an arm’s length arrangement rather than
foreign direct investment, developing countries do not need to use direct subsidies for
multinational ﬁrms. To encourage licensing, it is sufﬁcient for developing countries to
adjust contract enforceability, 6. If real plant-level ﬁxed costs incurred by foreign

direct investment are greater than critical value of ﬁxed costs (F > f ), then the ﬁrm

122

prefers licensing to foreign direct investment. On the contrary, if F < F but licensing

gives the host country greater social welfare, then the host country is likely to increase
contract enforceability (greater 6) to make F be lower than F since g—I;<O. An

increase in 6 induces the ﬁrm to choose an arm’s length arrangement rather than
foreign direct investment.

Here is more detail. If 7 > 1.5 in Figure 3.2, then the host country would decrease
6 as far as possible, but not so low 6 that the multinational ﬁrm will choose foreign
direct investment. It is because licensing gives more social welfare to the host country

than foreign direct investment (WI > Wf ), and a rise in 6 lowers the social welfare

. 6W,
under licensrng — < 0 .
66

Now consider 7 is located between 1 and 1.5. Speciﬁcally, suppose that 7:1.3.

Figure 3.4 in Appendix shows that the host country would choose 6:0.4 if the

multinational ﬁrm still chooses an arm’s length arrangement when 6 = 0.4 and 7 =1.3 .

The arrows in Figure 3.4 display directions of increase in the host country’s social
welfare.40 If the ﬁrm chooses foreign direct investment rather than licensing when the
host country sets 6 = 0.4, then the host country would increases 6 greater than 0.4 in
order for the ﬁrm to choose licensing since licensing gives greater welfare to the host
country than foreign direct investment. However, the host country sets the lowest value
of 6 under which the ﬁrm chooses licensing instead of foreign direct investment.4|

Turn to the case in which 7 is lower than or equal to 1. One example for that case

 

6W, 2 169(2 — 519)
a ——

‘0 When 7 = 1.3 , these directions are derived from — = (A- 3 . Thus, if 6 > 0.4,
4(13-50)

6W, 6W,
then —<0.0nthecontrary,if 6<0.4,then -—>O.
66 66

4‘ When I < 7 S 1.18 , numerical examples show that the host country’s social welfare is greatest at the
values of 6 on the dividing line (3 — 27 - 6 = 0 ).

123

7"

 

might be that a developed country (country M ) has some industries in which local
companies do not exist to compete with a multinational ﬁrm of the other developed
country (country N). Then, the country M can absorb easily new technologies of
country N . As previously stated, country M is likely to prefer foreign direct
investment to licensing (region B in Figure 3.2). If §< 7 <1 and no competent local
company exists, then the host country (country M ) would induce the multinational
ﬁrm of country N to choose foreign direct investment by changing 6. The country
M sets 6 to make F be equal to or greater than F because then the multinational
ﬁrm of country N chooses foreign direct investment.

Figure 3.5 in Appendix summarizes this section. Assume that the host country can
improve the capability to absorb new technologies (7) only in the long-run, but can
change the contract enforceability (6) in the short-run. Thus, the country changes 6 as
a policy instrument. Curve G (segmenting curve) is derived from equations (3.12) and
(3.13). In the left side of the Curve G (regions I and K), the host country prefers
foreign direct investment to licensing as the firm’s entry mode. But the country prefers
licensing in the right side of the Curve G (regions J and L). Curve H is derived
from equations (3.9) and (3.10). Above the Curve H (regions 1 and J), the
multinational ﬁrm prefers licensing to foreign direct investment. However, below the
Curve H (regions K and L), the ﬁrm prefers foreign direct investment.

From Figure 3.5 in Appendix, we can infer that the host country and the
multinational ﬁrm are in conﬂict over ﬁrm’s entry mode when parameters (6 and 7)
are located in regions I and L. However, when parameters are in regions J and K,
the host country’s interest is congruent with that of the ﬁrm.

If 7 is greater than the value on Point M where Curve H intersects the

dividing line (3-27-6=O), then the host country chooses contract enforceability

124

along the Curve H 42 depending on the real value of 7 since the host country’s social

welfare is greater under licensing (WI > W f) and an increase in contract enforceability

6W
decreases the country’s welfare [76% 0]. When 7 is greater than 1 but less than

the value on the Point M , then the dividing line (3—27—6 = 0) is the host country’s
optimal policy trace. In other words, the host country chooses 6 along the dividing
line when the values of 7 are in the aforementioned range.

If -2L< 7 <1, then the host country never chooses perfect contract enforceability. For
example, although the host country is very developed and thus has great capability to
absorb new technologies, it is possible that the country chooses imperfect contract
enforceability. This is because the host country prefers foreign direct investment to
licensing but perfect contractibility makes the multinational ﬁrm choose licensing
instead of foreign direct investment.

To extend the model, suppose that the host country prefers direct subsidies to
changing contract enforceability. If the host country needs many processes to change
contract enforceability, then the country might prefer direct subsidies to changing
contract enforceability. Recall that if real ﬁxed costs, F , are greater than F , then the
multinational ﬁrm would choose licensing rather than foreign direct investment. Again
analyze the case where -;-< 7 <1 (in Figure 3.2) and the host country prefers foreign
direct investment to licensing.

Denote S direct subsidies by the host country for the multinational firm to
encourage foreign direct investment. If the host country’s social welfare under foreign

direct investment is sufficiently greater than under licensing, then the country is likely

 

‘2 We trivially assume that the multinational ﬁrm chooses licensing if parameters are exactly located
along Curve H . Thus, the Curve H is the host country’s optimal policy trace when 7 is greater than

the value on the Point M .

125

to use direct subsidies. In other words, if W f - S > W] and F —S s F ,43 then the host

country can induce the multinational ﬁrm to choose foreign direct investment by
subsidizing the plant-level ﬁxed costs. Since the host country is likely to subsidize as
little as possible for the multinational ﬁrm to choose foreign direct investment, the
direct subsidies are equal to F — F .44

By using the minimum subsidies, we can obtain the condition under which the host

country directly subsidizes the ﬁxed costs.

Wf-Wl-F+F>O

Represent this condition with parameters

72(A-a)2(1-0)(47-3)(4r-9-1) > F
2(27-1)2<27-6)2

 

(3.16)

where 6e(0,1] and 7e(§,1).

If the condition of (3.16) is satisﬁed, then the host country’s policy of direct
subsidies can beneﬁt both the host country and the multinational ﬁrm. The host
country’s consumers are better off under foreign direct investment than under licensing.

Thus, if the host country gives greater weight to consumer surplus than to producer

surplus (,6 <-2L), then the possibility that the host country uses direct subsidies as a

policy tool to promote foreign direct investment increases.45

4. Case 2: One Competent Local Company

To examine the effects of market competition on the multinational ﬁrm’s choice

 

‘3 For simplicity, here we assume that if F = F , then the multinational ﬁrm chooses foreign direct
investment.

“ Since the multinational ﬁrm chooses foreign direct investment when F - S = F, we can derive the
minimum subsidies from this equation.

‘5 We do not consider this issue further.

126

 

between licensing and foreign direct investment, turn to the case in which one local
company in the host country produces variety i before the multinational ﬁrm of
variety i enters the host country. In the previous section in which no local company in
the host country can compete with the multinational ﬁrm, we assumed asymmetric
information between the multinational ﬁrm of variety i and local companies of variety
j (where i¢ j ). In this section, we assume asymmetric information between
producers of variety i (the multinational ﬁrm and incumbent local company) and
producers of variety j .46

Thus, if the multinational ﬁrm presents a licensing offer to a local company of
variety j , the local company would accept that offer without any consideration since
the offer satisﬁes the local company’s individual rationality constraints.“ However, if
the multinational ﬁrm tries to make licensing contracts with the incumbent company of
variety i, then the ﬁrm must make an acceptable offer for the incumbent company. In
other words, the proﬁt of the incumbent company (licensee) of variety i under
licensing contracts should be equal to or larger than that under the incumbent
company’s refusal to the ﬁrm’s offer of licensing contracts.

In this section, we maintain the four stage game analyzed above. If one local
company of variety i already exists in the host country, then the multinational ﬁrm of

variety i must consider the market competition before deciding between licensing and

 

4" We might assume that there is asymmetric information between the multinational ﬁrm and local
companies of all varieties in the host country, rather than between producers of variety i (the
multinational ﬁrm and incumbent local company) and variety j (other local companies whose products
are different from that of the multinational ﬁrm). However, it is more realistic to assume that the
incumbent company of variety i has better information on the multinational ﬁrm’s advanced technology
than other local companies of variety j (where i at j) do. The possible reason for this assumption is
that the incumbent company might engage in R&D to improve its technology and hence the company has
knowledge on the multinational ﬁrm’s technology. Thus, the multinational ﬁrm and the incumbent
company of variety 1' are assumed to be able to calculate ex ante the rent shares under an arm’s length
arrangement, but the local companies of variety j are assumed that they do not know ex ante the

licensee’s rent share.
‘7 The local company of variety j ‘s opportunity cost is zero.

127

foreign direct investment. If the multinational ﬁrm chooses an arm’s length arrangement
with the incumbent company, then it can exert monopoly power in the host country.

However, if the ﬁrm decides to enter the host country by foreign direct investment or

make licensing contracts with a new local company of variety j, then two ﬁrms, the
afﬁliate (or licensee) of the multinational ﬁrm and the incumbent company, compete

with each other.

4.1 Foreign Direct Investment

First, we study the situation in which the multinational ﬁrm chooses foreign direct
investment. We still assume that the initial marginal cost to produce variety i is a
for both the incumbent company and the multinational ﬁrm’s afﬁliate.48 However, the
multinational ﬁrm can decrease its afﬁliate’s marginal cost by transferring technology,

(a -t f ). The respective proﬁts of the multinational ﬁrm and the incumbent company

are given by
m _ _ _ m __1_ 2 _
Irf—(A Qf a+tf)qf 27tf F (3.17)
7:6 =(A—Q —a)qc (3.18)
f f f
where Q! is total output in the host country under foreign direct investment and
Q/ 2 ‘17 “If"

The ﬁrst order conditions and simple calculations give us respective outputs of the

multinational ﬁrm and the incumbent company.

 

‘8 We can extend this model by assuming that the marginal cost of a new entrant in the market such as the
multinational ﬁrm’s afﬁliate or licensee is it: (0 < A < l or 2 21) but the marginal cost of the
incumbent company is a . That is to say, the marginal cost for new entrant in the market is different from
a . However, this would make the analysis very complicated.

128

m:A-—a+2t c A—a—t

qf 3 ’ qf '- 3
Using the above, we can represent equations (3.17) and (3.18) by parameters and
technology transfer (t f ). In the third stage, the multinational ﬁrm chooses the optimal
amount of technology transfer.

t _—_ 4(A—a)

f 97-8 (3.19)

Then, respective outputs of the multinational firm and the incumbent company, and total

output under foreign direct investment are given by

 

 

m __ 3704‘“)
qf ———97—8 (3.20)

c _ (A-a)(37-4)
qf — 97-8 (3.21)
Qf = 2(A—a)(37-2) (322)

97—8

From equations (3.19) to (3.22),49 we can ﬁnd the effects of change in the efficiency
of technology transfer (7) on the amount of technology transfer and outputs. If the host
country increases its capability to absorb advanced technologies (lower 7), the amount
of technology transfer and the multinational ﬁrm’s output rise, but the incumbent
company’s output decreases. Since the increase in the multinational ﬁrm’s output
outweighs the decrease in the incumbent company’s output, the total output in the host
country rises as 7 decreases.50

The respective proﬁt functions for the multinational ﬁrm and the incumbent

company are represented by parameters.

 

‘9 For a duopoly to be sustained in the market, assume that 7 > 4 / 3 in this section.

at _ 6 A- aq'" -24 A— We 12 A— 60 -12 A-
50 _f.=_3_<__g><o, __f_=__<_~j_><0, _f_= < $0,,“ _f:__<_§>
6r (97 - 8) 67 (97 - 8) 67 (97 - 8) 67 (9r - 8)

 

<0.

129

’l‘

m_7(A—a)2_

zf .. (974) (3.23)

c ___ (374904—01)2
(97-3)?-

 

” (3.24)

When differentiating equations (3.23) and (3.24) with respect to 7, then we obtain

m C
5’? -8(A-a)2 6”)" 24(A-a)2(37-4)
= 2 <0, and = 3
67 (97—8) 67 (97—8)

 

 

> 0. That is, if the incumbent

company competes with the multinational ﬁrm in the host country, a rise in host
country’s capability to absorb the ﬁrm’s advanced technology increases the
multinational firm’s proﬁt, but decreases the incumbent company’s proﬁt under foreign

direct investment.

Q2
The host country’s welfare is made up of consumer surplus [% and the

incumbent company’s proﬁt (7r; ). Thus, the host country’s social welfare under foreign

direct investment becomes

3(A—a)2(972 —167+8)
W f = 2
(97-8)

 

(3.25)

From equation (3.25), we can derive that an increase in the host country’s capability to

 

6W _ _ 2
absorb the technology raises the country’s welfare, f = 48(A a3) < 0, although
57 (97 - 8)

it decreases the local company’s proﬁt.
Let us analyze the'situation in which the multinational ﬁrm chooses an arm’s length
arrangement. We divide the situation into two sub—cases, one in which the host country

prevents the multinational ﬁrm from making licensing contracts with the incumbent

130

company of variety i because of some economic reasons such as to inhibit the
multinational ﬁrm from extracting monopoly proﬁts, and the other in which the host
country allows the multinational ﬁrm to make licensing contracts with the incumbent
company. We begin with the sub-case in which the host country sets direct restrictions

on any licensing contracts between the multinational ﬁrm and the incumbent company.

4.2 Contract with a New Local Company

We examine the sub-case in which the host country restricts licensing contracts
between the multinational ﬁrm of variety i and the incumbent company of variety i.
The host country might inhibit the multinational ﬁrm of foreign country from making
licensing contracts with the monopolistic local company because of concern about
outﬂows of monopoly proﬁts. Thus, the multinational ﬁrm of variety i would make
license contracts with a local company of variety j (where i: j). Since the licensee
of the multinational ﬁrm competes with the incumbent company, the host country’s
market structure results in a duopoly. The respective proﬁts of the multinational ﬁrm

and the incumbent company are

m_ m 1 2
7r”, —- 6(A—a—in +511)an —§7tnl (3.26)

”:1 =(A—a-Qn,)q;‘, (3.27)

where subscript n] denotes licensing contracts between the multinational ﬁrm and a

new local company, 71:] is the incumbent company’s proﬁt, and Q”, = qr] + quiI‘

Two ﬁrst order conditions and simple calculations give respective outputs of the

multinational ﬁrm and incumbent company as a function of parameters and technology

A—a+2t A-a—t
transfer, such as q”; =—————i, and qul =—?—'i. In the third stage, the

3

multinational ﬁrm chooses the extent of technology transfer

131

= 46(A—a) (3 28)
"I 97—86 '

Stricter contract enforceability induces the multinational ﬁrm to transfer more

$1 = 367(A—a)

technology, 66 (97-86)2

> 0. From equations (3.19) and (3.28), we ﬁnd that the

multinational ﬁrm transfers less technology under licensing contracts with a new
company than under foreign direct investment.
Using equation (3.28), we can obtain respective outputs of the multinational ﬁrm

and the incumbent company, and total output

m _ 3701-0)

 

 

(In, - 97—86 (3.29)
c _ (A—a)(37—46)
Q", = 2(A—a)(37—26) (3.31)

97—86

Both an increase in contract enforceability and an improvement in the capability to

absorb the technology increase total output in the market.

 

 

aQ,,, ___127(A-a)>0 and 6Q", :—1219(A-a)<0
66 (9),—1319)2 07 (97-86)2

From equations (3.22) and (3.31), we ﬁnd that total output under foreign direct

investment (Q f) is not less than under an arm’s length arrangement with a new local

company (in ). Thus, although the market results in a duopoly for aforementioned both

entry modes, the mode of foreign direct investment makes the consumers better off.
F urtherrnore, an improvement in the host country’s adaptability to the ﬁrm’s technology

increases total output under foreign direct investment more rapidly than that under

132

licensing contracts with a new local company.

6(Qf —Qn1): 12(A—a)(1-6)(646—8172)

2 2 5°
57 (97-8) (97-86)

 

 

The respective proﬁts of the multinational ﬁrm, the incumbent company, and the

licensee in the host country become

 

 

2
m 9704-0)
=———— 3.32
””1 97—86 ( )
2 2
c _ (A-a) (37-49)
It", — 2 (3.33)
(97-86’)
_ 2 _ 2
n2]: 9“ (9)7 (’4 2“) (3.34)
(97-89)
where 7:21 is the licensee’s proﬁt (2,121 = (l—6)(A—a—in +tn1)qz;). It shows that

an increase in contract enforceability raise the multinational ﬁrm’s proﬁt but decreases
the incumbent company’s proﬁt. An improvement in the host country’s adaptability to
the ﬁrm’s technology increases the multinational firm’s proﬁt, but decreases the
incumbent company’s proﬁt.5 '

Although licensee’s proﬁt is weakly increasing in host country’s absorptive capacity

6751,, : —144719(1 — 6)(A — a)2
67 07—8603

 

 

regardless of parameter values 3 0 , the effects of an

increase in contract enforceability on the licensee’s proﬁt depend on parameters

 

m 2 2 c
aI‘rrrl=97 (”1'“) >0 anti]:

 

 

5' From equations (3.32), (3.33), and (3.34), we can derive that

 

 

 

 

as (97 — 139)2 ’ aa
—247(A -a)2(37 — 49) < 0 Buff} _ —802(A -a)2 < 0 an; _ 246(A -a)2(37 — 40) > 0
(97 — 86)’ ’ ar (97 - 86)’ ’ ar (97 - so?

133

I

67: 2 _ 2 _ _

4:97 (A a) (16 397 86) . If 7>1§,ti , then an increase in contract
66 (97-86)

 

enforceability decreases the licensee’s proﬁts irrespective of 6. However, if %< 7 < L:- ,
then an increase in contract enforceability makes the licensee better off. This implies
that if the multinational ﬁrm can transfer its technology efficiently (lg) < 7 < % ), then the
multinational ﬁrm might transfer technology to a greater extent to the host country
because of its greater share of operating proﬁts. In other words, although the licensee’s
fraction of operating proﬁts decreases due to an increase in contract enforceability (6),
the licensee’s proﬁt increases because the multinational ﬁrm transfer a greater extent of
technology and hence the operating proﬁts grow rapidly.

Using equations (3.23) and (3.32), we can also ﬁnd a critical value of plant-level
ﬁxed costs, F , such that for F S F , the multinational ﬁrm chooses foreign direct

investment as an entry mode.

972(A-a)2(1—6)
(97—8)(97-86)

F:

 

(3.35)

Similar to section 3, we can conﬁrm that the critical value of ﬁxed costs decreases with

respect to 6, and weakly decreases with respect to 7.

_ai __ -972(A-a)2 < 0
66’ (97—80)’

 

f = 727(A-a)2(l-6)(166—97—976) < o
07 (974)267-8602 ’

 

The range of the critical value is given by

_ 2
03F<M2Fm (3.36)
97-8

134

where F m is the critical value when 6 = 0.

The host country’s producer surplus consists of the proﬁts of two local companies,

the incumbent company and the licensee. Hence, consumer surplus [Q—Z’Z’L) and the

proﬁts of two local companies (2:! + ”1111) form the host country’s social welfare.

W : 3(A—a)2(1272 —3726—1676+862)

 

(3.3 7)

We can conﬁrm that an improvement in capability to absorb the technology increases

 

6W _ 2 _ ..
the host country’s welfare, ——”—1- = 48“ a) 98793 6 37) < 0.
57 (97-86)
If 7 > 8 46+4 9 0+4 , then an increase in contract enforceability decreases

 

9

6W"; = 3(A ‘0')27(-2772 + 487 ~2476 +166)

 

 

 

the host country’s welfare, 3 < 0. That is,
(97 - 86)
_ ,/ 2 _
if the ﬁrm cannot transfer its technology efficiently [7 > 8 40 + 4 96 0 + 4 J , then a

decrease in contract enforceability makes the host country better off. However, a

decrease in contract enforceability increases the possibility that the multinational ﬁrm

chooses foreign direct investment instead of licensing [6% < 0).

Comparing equation (3.37) with (3.25), we can show that the host country’s social
welfare under licensing contracts with a new local company is greater than under
foreign direct investment except when both 7 and 6 are very small (see Appendix 2).
Unless the multinational can transfer its technology very efﬁciently, then the host
country prefers licensing contracts with a new company to foreign direct investment.

Consumers are not worse off under foreign direct investment than under licensing

135

contracts with a new company (Q f 2 in). But producer surplus in the host country is

larger under licensing contracts with a new local company (7!; < 2:] + 2,111 ).52

The host country sets contract enforceability, 6, to induce the multinational ﬁrm to
choose an entry mode desirable to the country. If the multinational ﬁrm cannot transfer
technology efﬁciently, then the host country is inclined to set 6 to make F > F for
the multinational ﬁrm to choose licensing contracts with a new company. But if the ﬁrm
can transfer its technology efﬁciently (low 7), then the host country might prefer
foreign direct investment as an entry mode since the gain in consumer surplus could
outweigh the loss of producer surplus. Similar to the previous section, assume the host
country prefers direct subsidies to changing contract enforceability in the short-mn.
Then, it is possible for the country to subsidize the multinational ﬁrm to encourage
foreign direct investment if subtracting subsidies from social welfare under foreign

direct investment is greater than social welfare under licensing (W f — S > W", ).

4.3 Contract with the Incumbent Local Company

Now turn to the sub-case in which the host country allows the multinational ﬁrm of
variety i to make licensing contracts with the incumbent company of variety 1'. When
the multinational ﬁrm chooses licensing rather than foreign direct investment, it can
exert monopoly power in the host country. However, to induce the incumbent company
to make licensing contracts with the multinational ﬁrm, the ﬁrm must present the
incumbent company with an acceptable offer. The incumbent company would accept

any offers to guarantee its proﬁt which is equal to or larger than that under direct entry

 

,, c c , 37(A—a)2(l-6)(24076—2566—24373+487+21672)
7’] _’7nl __ﬂnl = 2 2
(97-8) (97-89)

 

<0, where 7>§ and
6e(0.1].

136

(foreign direct investment) by the multinational ﬁrm.”

Hence, any acceptable offer must satisfy the following constraint
C
(l—6)(A—qil—a+til)ql.l +h27rf (3.38)
where subscript i1 denotes licensing contracts between the multinational ﬁrm and the
incumbent company, and (l — 6)( A — q” — a + (1.1)qu is the incumbent company’s share

of operating proﬁts when the incumbent company makes licensing contracts with the
multinational ﬁrm,54 and h denotes the multinational ﬁrm’s offer. Thus, the minimum

ﬁrrn’s offer to the incumbent company is

h=7rj,—(1-6)(A—qi1—a+til)qil (3.39)

When the multinational ﬁrm makes an acceptable offer to the incumbent company,

the ﬁrm’s proﬁt becomes

 

5’ The multinational ﬁrm of variety i can enter the host country by making licensing contracts with a
new local company of variety j (where irt j ) instead of foreign direct investment if the incumbent

company of variety 1' do not accept the multinational ﬁrm’s offer. In this case, the multinational ﬁrm
must make an offer to guarantee the incumbent company’s proﬁt that is equal to or larger than the proﬁt
under licensing contracts with a new local company if the firm wants to make licensing contracts with the
incumbent company. However, the incumbent company does not know ex ante which mode the
multinational firm will choose if the incumbent company denies the multinational ﬁrm’s offer. Comparing
equation (24) with equation (33), we can ﬁnd that the proﬁts of the incumbent company under foreign
direct investment are less than under the multinational ﬁrm’s licensing contracts with a new local

company (7:; < ”:1 , see Appendix 4.3.1). If accepting the offer makes the incumbent company be worse

off than under foreign direct investment, then the incumbent company always denies the offer. In addition,
the focuses lie in comparison between the host country’s welfare under foreign direct investment and that
under licensing contracts with the incumbent company, and comparison between welfare under foreign
direct investment and that under licensing contracts with a new local company. Hence, we assume that the
incumbent company would accept any offers to guarantee the proﬁt that is equal to or larger than that
under foreign direct investment.

5" We assumed that the incumbent company of variety 1' knows ex ante the shares of operating proﬁts.
but local companies of variety j do not know ex ante.

137

m _ 1 2
”i1 ‘ 9(A’qil‘“+’i1)qil‘57’i1‘h

l
l
= (’4 ‘qu ”“16qu “3702 ‘7’;

Surprisingly, if the multinational ﬁrm can make licensing contracts with the incumbent
company, contract enforceability in the host country does not affect the ﬁrm’s proﬁts.
Hence, if the incumbent company accepts the multinational ﬁrm’s offer, the

multinational ﬁrm would not consider contract enforceability in the host country.

A-a+L
In the last stage, the ﬁrm chooses the output of variety i, q” = —2—-’1-. In the

third stage, the ﬁrm chooses the extent of technology transfer

t. :11 (3.41)
’1 27—1

Inserting equation (3.41) into the output formula, then we obtain the multinational

ﬁrm’s output

_Mﬂ

. — 3.42
qr] 27_1 ( )

Interestingly, when 7 < 2 , total output under licensing contracts with a new local
company (in) could be less than under licensing contracts with the incumbent
company (qu ).55 That is, if the ﬁrm’s technology can be transferred efﬁciently

(1;- < 7 < 2) and the contracts are poorly enforced in the host country (low 6), then the

monopoly output (qu ) could be greater than duopoly output (in) in the market.

When comparing equation (3.41) with (3.19), we can show that the multinational

 

55

_ (A -a)(372 -67 + 40)
(2r-IX97-80) '

 

in — qil

138

ﬁrm transfers greater amount of technology under licensing contracts with the
incumbent company than under foreign direct investment if and only if 7 > 4 .56 From
equations (3.20) and (3.42), we also ﬁnd that if %< 7 <4}, then the multinational firm
produces more output under foreign direct investment than under licensing contracts
with the incumbent company.57 However, irrespective of the range of 7 , total output in
the market is greater under foreign direct investment than under licensing contracts with
the incumbent company (equations (3.22) and (3.42)).58

The multinational ﬁrm’s proﬁt is

”m _ (37—2)(1572 —I6)(A—a)2
’1 2(27—l)(97—8)2

 

(3.43)

If the host country permits the multinational ﬁrm to contract with the incumbent
company, the multinational ﬁrm always prefers an arm’s length arrangement to foreign
direct investment (see Appendix 4). This is because the multinational ﬁrm can sustain
its monopoly power in the market under licensing contracts with the incumbent
company and extract all the proﬁts generated by technology transfer. But if the ﬁrm
chooses foreign direct investment, it must share the market with the incumbent company.

The host country’s social welfare under licensing contracts with the incumbent

2
C q. I
company consrsts of consumer surplus [fl] and the 1ncumbent company’s proﬁt

under foreign direct investment (It; ). Thus, the country’s welfare is given by

 

t —:. = (A-ax4—7)
f " (9r-8X2y—I)’
=74A-aX5-bd
(97-8)(27-l)
2
n (A-aXbr-67+4)
Q -q. = >0.
f " (27-1)(97-8)

56

 

57 m
qf 'qu

 

139

 

W _ (A —a)2(15374 —40873 + 40272 —1767 +32)

. —- 3.44
” 2(97-8)2(27-1)2 ( )

Comparing equation (3.44) with equation (3.25), we ﬁnd that foreign direct investment
gives greater welfare to the host country than licensing contracts (see Appendix 5).
Knowing that social welfare under licensing contracts with the incumbent company
is smaller than under foreign direct investment, the host country tends to restrict the
licensing contracts with the incumbent company when only one competent local
company produces the same variety to that of multinational ﬁrm. The host country also
knows that the incumbent company would not accept the multinational ﬁrm’s offer if
the offer is negative (that is, h < 0 ).59 In other words, if the multinational ﬁrm makes a
minus offer (h < O) to the incumbent company, the incumbent company would refuse
the offer and then can make more proﬁts. Thus, for the incumbent company to refuse

the multinational ﬁrm’s minus offer, the host country chooses the level of contract

enforceability in the range from 0 to 6 , where 7r; = (l - 6)(A _qu —a Hi] ”6' From

equations (3.24) and (3.39), we can illustrate the host country’s choice of 6 in {7,6}
space (see Appendix 6).

In Figure 3.3, the curve for h = O separates region of the multinational firm’s
positive offers (B) from region of negative offers (C ). The host country knows that the
incumbent company never accepts the multinational ﬁrm’s negative offer (h< 0) and
the country’s welfares are greater under foreign direct investment than licensing
contracts with the incumbent company. Thus, the host country always chooses contract

enforceability below 6. It is implied that region (B) in Figure 3.3 is not chosen by the

host country. In other words, perfect contract enforceability is never chosen if one

 

’9 From equation (3.39), h < o if It; >(1-6)(A—ql.l -a “11)‘111'

140

competent local company already produces variety L60 Thus, the host country does not
need to prohibit directly the multinational ﬁrm from making licensing contracts with the
incumbent company‘sl Instead, the host country can use a policy instrument, 6, to
acquire same effects to direct restrictions on licensing contracts between the

multinational ﬁrm and the incumbent company.

Figure 3.3: Region of Parameters 6 and 7 for the Host Country’s Choice of 6

"-- l l
1 -'s h>0 (B)

-
"s
‘---

 

h 2'0 Curve

""" h<0 (C)

 

 

 

1.3 y 6

In conclusion, if one local company in the host country can compete with the

multinational ﬁrm, then the country chooses contract enforceability (6) in the range of
(0,6) for the incumbent company to refuse the multinational ﬁnn’s offer. In addition,

if the host country prefers licensing contracts with a new local company of j to

 

6° In this section, assume that 7 >% for duopoly to be maintained in the market. When 7 is 1.3, then

6 is 0.999. As we see in Figure 3.3, the greater 7 is, the smaller 6 is.
6' However, the multinational firm can collude with the incumbent company and share the monopoly

profits with the incumbent company although the host country sets 6 smaller than 6 . We exclude the
possibility of collusions between the multinational ﬁrm and the incumbent company because the host
country would prohibit such collusions.

I41

foreign direct investment, then the country would set 6 for the ﬁrm to choose

licensing (F >F )62 This section shows that if one incumbent local company is

competitive, then the host country never chooses perfect contract enforceability.

5. Case 3: Two Competent Local Companies

What about a case in which two competent local companies in the host country
produce the variety i? If two competent local companies produce variety i before the
multinational ﬁrrn’s entry into the host country, then the multinational ﬁrm might have
less incentive to enter the host country than the cases 1 and 2, as Blomstrom & Kokko
(1997), and Neary (2002) showed};3 We still assume asymmetric information between
producers of variety i and producers of variety j (where i¢ j).

In the previous section, the host country sets 6 to inhibit the multinational ﬁrm
from making licensing contracts with an incumbent company. In contrast to the previous
section, the host country has little reasonable motive to prevent the multinational ﬁrm
from making licensing contracts with one of the incumbent companies since the two
incumbent companies produce variety i in this section. However, the host country
would have a strong incentive to inhibit multi-licensing contractsﬁ4 As Shapiro (1985)
emphasized, the multinational ﬁrm can design and implement perfectly the monopoly
(fully collusive) outcome by using multi-licensing contracts. By using multi-licensing
contracts, the multinational ﬁrm can extract the proﬁts originated by monopolization.
Thus, the host country would be worse off under multi-licensing contracts than under
any other entry modes such as single-licensing contracts and foreign direct investment.

Thus, we assume that the multinational ﬁrm can make licensing contracts with one

 

‘2 However, 6 should be located inside region C in Figure 3.3.

63 In other words, an increase in the number of incumbent companies of variety 1' in the host country
decreases the incentive of the multinational ﬁrm of variety 1' to enter the host country.

6‘ Multi-licensing wntracts mean that the multinational ﬁrm makes licensing contracts with two
incumbent companies of variety 1' .

142

of the incumbent companies. If neither of the incumbent companies accepts making
contracts with the multinational ﬁrm, then the ﬁrm has no choice but to enter the host
country by foreign direct investment or by licensing contracts with a new local company
of variety j . However, if one of the incumbent companies does not accept the
multinational ﬁrm’s offer, then the other incumbent company can accept the ﬁrm’s offer.
C1 and C2 denote the incumbent companies of variety 1', 1 and 2 respectively. For
simplicity, we assume that the multinational ﬁrm makes licensing contracts with C1. If
C l reﬁises the multinational ﬁrm’s offer, then the multinational ﬁrm would make
licensing contracts with C2. Thus, C1 would accept any offers that would satisfy the

following condition

(1 —6)(A—Q[ —a+t1)q[m +112 22le (3.45)

where 7:16;; represents C I ’s proﬁt under licensing contracts between the multinational

ﬁrm and C2. In addition, q,’" is output of the ﬁrm’s licensee (Cl) and

Q1=q1m +qlC2, where qlc2 is output of C2 under licensing contracts between the

multinational ﬁrm and C 1.

In contrast to the previous section, the multinational ﬁrm’s offer h can be negative
(h < 0) in this section since the two incumbent companies would bid to be a licensee of
the multinational ﬁrm. The two incumbent companies would earn identical proﬁts due

to their biddings to be a licensee.65 Hence, the multinational ﬁrm’s offer becomes
h = 7:10 -(1—6)(A -Q, -a +111)qu

The multinational ﬁrm’s proﬁt is

 

6’ ”(all = 7:3 = 7:3 = nlclz = 7r,“ where superscript C1 (C2) represents the local company Cl (C 2 )

and subscript ll (12 ) represents licensing contracts between the multinational ﬁrm and C 1 (C2 ).

143

1
ﬁlm =6(A—Ql—a+t1)qlm —§7t12—h
1
= 6(A—Q1—a+tl)q1’" 77:17- —;zf +(l—6)(A—QI -a+z,)q,’" (3.46)

1
=(A-Q,—oz+t,)q,"’--2-7r,2-(A—Q,-oz)qf2

Equation (3.46) shows that contract enforceability does not play any role when two

incumbent companies compete to be a licensee of the multinational ﬁrm. The proﬁt of

C2 is
zﬁ’ = rrf =(A—Q, —a)qf2

In the ﬁnal stage, the multinational ﬁrm and C2 choose their outputs. The ﬁrst

order conditions and simple calculations give respective outputs of the multinational
2

ﬁrm and C2; qu= , and q, . In the third stage, the

multinational ﬁrm chooses the extent of technology transfer

I = 5(A—a)

3.47
1 87—5 ( )

We can easily compute the outputs of the multinational ﬁrm (licensee) and C2 66

m : 4704-0!)

 

 

1 87 _ 5 (3.43)
c2 _ (A —a)(47 -5)

q, — 2(87 _ 5) (3.49)
_ (A-aX127-5)

Q, — 2(87 _ 5) (3.50)

We again conﬁrm that the host country’s improvement of the adaptability to the ﬁrm’s

 

6" In this section, for the incumbent companies of variety i to produce under the entry of the
multinational ﬁrm in the host country, we assume that 7>§ (see equation (3.57)).

144

6Q, = —10(A -a)
67 (87 —5)2

 

technology increases total output [ <0], but decreases the C 2 ’s

 

6qu _10(A—a)>0]

output
[ 67 (87—5)2

The respective proﬁts of the multinational ﬁrm and C2 are represented by

”m = (27+5)(A-a)2

 

 

1 4(8), _5) (3.51)
2 2
7:102 = (49:5) (”1‘2“) (3.52)
(87- 5)

Similarly, an increase in capability to absorb the ﬁrm’s technology in the host country
raises the multinational ﬁrm’s proﬁt, but lowers the C2 ’5 proﬁt.67

If the multinational ﬁrm enters the host country by foreign direct investment, the
ﬁrm must compete with two identical incumbent companies, C1 and C2. The

respective proﬁts of the multinational ﬁrm and the two incumbent companies are

1
n7=(A-Qf—a+zf)q’f"—Ey:}—F (3.53)
7r;1=7r;2=7r;=(A—Qf—a)q;i, i=1,2 (3.54)

where Q f = q’f" + q? + q? = q’f" + 2g;i (i =1, 2 ). We can represent respective outputs

of the multinational ﬁrm and the incumbent companies as a function of parameters and
4

4

c2_

technology transfer, such as q? = and q"1 =q f — . The

extent of technology transfer under foreign direct investment is

 

.7 an,“ -256: -a)2 87,” 10(A -a)2(47 — 5)
= 2 < 0 , and = 3 >
67 2(87 -— 5) 57 (87 - 5)

 

 

 

0.

I45

= 3(A-a)

if 87—9 (3.55)

Comparing equation (3.55) with equation (3.47), we ﬁnd that the multinational ﬁrm
transfers technology to a greater extent under foreign direct investment than under an

arm’s length arrangement<58 only when the ﬁrm can transfer its technology to the host
country efﬁciently (%- < 7 < 183). But if the ﬁrm cannot transfer its technology efﬁciently
( 7 >181), then the ﬁrm transfers more technology under licensing than under foreign
direct investment.

Using equation (3.55), we can calculate respective outputs of the multinational ﬁrm

and the incumbent companies.

 

 

m _ 2704‘“)
qf -_—87—9 (3.56)
cl _ c2 _ (A‘aX27’3)
qf ‘4f — 87—9 (3.57)
_ 6(7—1XA-a)
Qf — 87-9 (3.58)

The total output is greater under foreign direct investment than under licensing‘s9 only

when the technology can be transferred efﬁciently (%< 7<185). An increase in

efﬁciency of technology transfer obviously induces the multinational ﬁrm to transfer

 

6t
_2 A-
technology to a greater extent [ f - 4( a)

— <0 under foreign direct
(67 (87—9)2 ]

investment. In addition, an increase in the host country’s adaptability to the ﬁrm’s

technology results in greater total output in the market although the local companies,

 

_ 2(A-a)(87-15)
f (87-5X8r-9) '
_ _ (A-a)(87-15)

f 2(37-5x37-9)°

 

68
t1-

 

69 Q,

146

C 1 and C2 , decrease their outputs.70

The respective proﬁts of the multinational ﬁrm and two incumbent companies are

_ 2

 

zrf — 2(87—9) (3.59)
2 2
f f 187—9)2

We can still conﬁrm that an increase in the host country’s capability to absorb the
multinational ﬁrm’s technology raises the multinational ﬁrm’s proﬁt, but decreases the
incumbent companies’ proﬁts.“ From equations (3.59) and (3.51), we obtain a critical

value of ﬁxed costs, F , like sections 3 and 4.

(A -a)2(45—327)
4(87-9)(87-5)

F:

 

(3.61)

However, under the assumption of 7>% , F is always negative. Thus, the

multinational ﬁrm always prefers licensing to foreign direct investment when two
incumbent companies operate in the host country. 72
In the ﬁrst stage, the host country must compare two welfares, between licensing

and foreign direct investment. When the multinational ﬁrm chooses licensing, then the

host country’s social welfare consists of consumer surplus [+) and two local

companies’ proﬁts (27716 ). Similarly, social welfare under foreign direct investment

 

 

 

 

6Q _ aqcl (3ch _ 6qm _ _
70 f=—6(A a)<0 f = f =6(A a)>0,and f = 18(A a)<

67 (87-9)2 , 67 67 (87-9)2 67 (87-9)2
m 2 Ci 2
,, 57'; _—9(A—a) (mm 5’7 =12(A-a) (27—3)

67 2(37 — 9)2 67 (8r - 9)3
72 This is because if the multinational ﬁrm decides on licensing, then the firm does not need to pay ﬁxed
costs. Furthermore, licensing does not increase market competition.

0.

 

 

>0,where i=l,2.

I47

2

Q
consists of consumer surplus (7f) and producer surplus (22; ). The respective host

country’s social welfares under licensing and foreign direct investment are

_ (A —a)2(20872 — 2807 +125)

 

 

W 3.62

’ 8(87—5)2 ( )
2 2

Wf : 2(A—a) (137 3307+18) (3.63)
(87-9)

Figure 3.10 in Appendix shows which mode is better for the host country, licensing

or foreign direct investment; the answer depends on the host country’s capability to

absorb the technology, 7 (see Appendix 7). If 7 is in range C (% < 7 < 3.34) , then

the host country prefers an arm’s length arrangement to foreign direct investment. When
7 is located in range B G < 7 < 15) or D (3.34 < 7) , then foreign direct

investment gives greater welfare to the host country than licensing. The results imply
that, in general, the host country should be skeptical about claims that promoting
foreign direct investment will always raise its welfare when two competent local
companies operate in the host country as Hanson (2001) argued. This section shows that
the claim that promoting inward foreign direct investment raises the country’s welfare

might be true only when 7 are in the speciﬁc ranges.

6. Conclusion

We present a simple model to examine the effects of contract enforceability, plant-
level ﬁxed costs, and market structures on the multinational ﬁrm’s choice between
licensing and foreign direct investment. The multinational ﬁrm’s choice of entry mode

inﬂuences the host country’s social welfare. Thus, the host country tends to use various

148

policy instruments to induce the multinational ﬁrm to choose a desirable enhy mode.
Three cases considered are when the host country has: (1) no incumbent local company
that can compete with the multinational ﬁrm, (2) one incumbent local company, and (3)
two incumbent local companies.

The model explains why developing countries usually prefer licensing to foreign
direct investment in the early stage of the development. It is highly probable that
developing countries do not have any local companies to compete with multinational
ﬁrms of developed countries and would have difficulties in absorbing the advanced
technologies of multinational ﬁrms (high 7 ). This paper shows that developing
countries can change contract enforceability to encourage the multinational ﬁrm to
choose an arm’s length arrangement.

If a developed country with greater capability to absorb new technologies (low 7)
has some industries in which no incumbent local company can compete with
multinational ﬁrms, then the country would prefer foreign direct investment to licensing.
Thus, a host country is likely to choose the optimal contract enforceability based on its
capability to absorb new technologies (7 ). We show that if a host country prefers direct
subsidies to changing contract enforceability as a policy instrument, then the country
would subsidize the multinational ﬁrm to encourage foreign direct investment under
certain conditions. The direct subsidies could beneﬁt both the host country and the
multinational ﬁrm.

The host country never chooses perfect contract enforceability if the country has one
local company that can compete with a multinational ﬁrm. The multinational ﬁrm has a
strong incentive to make licensing contracts with the incumbent company to maintain
the monopoly position. However, the host country’s social welfare under licensing

contracts with the incumbent company is lower than under foreign direct investment.73

 

7’ For example, Gilbert and Newbery (1982) demonstrated that a ﬁrm with monopoly power could have

149

 

We show that the host country should choose imperfect contract enforceability for the
incumbent company to refuse the multinational ﬁrm’s negative offer (h < 0 ).

If the host country has two incumbent local companies to compete with the
multinational ﬁrm, then the country would use direct restrictions on multi-licensing
contracts. The multinational ﬁrm always prefers licensing to foreign direct investment
since foreign direct investment increases the competition in the market. However, the
host country’s preference for multinational ﬁrm’s entry mode depends on the country’s
capability to absorb the ﬁrm’s technology.

To maximize social welfare, the host country can choose among the various policy
instruments such as contract enforceability, direct subsidies, and direct restrictions. The
host country tends to adjust its policy by reacting optimally to its capability to absorb
advanced technologies and market conditions the country faces. The real value of 7 is
dependent of both the developing stage of the country and the characteristics of the
industry. The actual policy instrument used by the country is likely to vary between
industries, depending on the capability to absorb advanced technologies.

A developing country is likely to have some industries that can easily absorb
advanced technologies (low 7) although no incumbent local company is there to
compete with multinational ﬁrms in the industries. Then, the developing country might
set up policies to encourage foreign direct investment due to those industries. If a
developed country has some industries that would have difﬁculties in absorbing new
technologies (high 7 ) and no local incumbent company can compete foreign
multinational ﬁrms in those industries, then the developed country might induce

multinational ﬁrms to choose licensing in those industries.

 

an incentive to preserve its monopoly power by preemption. They showed that the ﬁrm’s preemptive
activity to maintain its monopoly power would be undesirable since the ﬁrm may spend resources on the
research and development, and then deny society the use of its new invention.

150

 

APPENDIX

1. Without Competent Local Company

Figure 3.4: Maximum Point of Social Welfare under 7 = 1.3

 

 

 

 

1
{h
l
1
. Direction of Increase in Social Welfare
FBI '0 2 Licensing
(B) I. (C)
L— I
9(7) 1
"’02 .
' (e)
I
\
0 : 6W, / 66 < 0

 

 

05

 

 

 

 

 

151

 

2. Comparison of Social Welfares
Subtracting equation (3.25) from equation (3.37) gives the following

37(A —a)2(l — 19)(24373 — 43272 +1207 — 7276+ 1286)

(A.1)
(97-86)2(97—s)2

 

G denotes the above equation (G s Wnl — W f ). A positive sign of G means that the

host country’s welfare under licensing contracts with a new local company is greater
than under foreign direct investment. The sign of G is identical to that of
(24373 —43272 +1207-7276+1286) . Deﬁne f(7,6)s(24373 —43272 +1207
—7276+1286). Figure 3.6 illustrates the function f (7, 6) in {7,6} space. Figure 3.6
shows that for only low values of 6 and 7, social welfare under licensing is smaller

than under foreign direct investment. Table 3.1 is obtained with a few numerical values

of6and 7.

Figure 3.6: Function f (r,6) for Comparison of Social Welfares

 

f

310‘-

 

 

 

 

 

 

152

Table 3.1: The Values of f

 

 

7 = 1.35 7 = 2 7 = 3
6 = 0.1 -24.4 454.4 3024.2
6 = 0.5 -12.0 448.0 2989.0
6 = 0.9 0.27 441.6 2953.8

 

3. Comparison of the Incumbent Company of Variety i ’s Profits

When we subtract equation (3.33) from equation (3.24), we obtain the following

 

247(A —a)2(1—6)(—2772 +307 +3076-326)

A.2
(97—86)2(97—8)’ ( )

Y denotes the above equation (Y E 7t;-II;1). A negative sign of Y means that the

proﬁt of the incumbent company of variety i under foreign direct investment is
smaller than under the multinational ﬁrm’s licensing contracts with a new local

company. Similar to Appendix 2, the sign of Y depends on the sign of
(—2772+307+3076—326). We again deﬁne g(7,6)a(—2772+307+3076—326).
Figure 3.7 illustrates the function g(7, 6) in {7,6} space. The Figure shows that the

proﬁts of the incumbent company are always smaller under foreign direct investment
than under licensing contracts with a local company of variety j . Table 3.2 is obtained

with a few numerical values of 6 and 7.

153

Figure 3.7: Function g(r,6) for Comparison of Profits

 

 

 

 

 

Table 3.2: The Values of g

 

7 = 1.35 7 = 2 7 = 3
6 = 0.1 -7.86 -45.2 -147.2
6 = 0.5 -4.46 -34.0 -124.0
6 = 0.9 -l .06 -22.8 -100.8

 

4. Multinational Firm’s Choice of Entry Mode

In order to obtain a critical value of plant-level ﬁxed costs, F , use equations (3.23)

and (3.43).

(A —a)2(—973 —2072 +647—32)

2 (A3)
2(97-8) (27-1)

F:

 

A negative sign of F means that the multinational ﬁrm always prefers licensing
contracts with the incumbent company to foreign direct investment irrespective of real
plant-level ﬁxed costs since we assume that the real plant-level ﬁxed costs are positive.

154

The sign of F is determined by (—973—2072+647—32). We also deﬁne the

function v(7)E—973 -2072 +647—32. The function v(7) is always negative under

57(7)

a =—2772—407+64<0and v(l.3)=—2.37.
7

7 >4 since

Figure 3.8: The Function of v( 7)

 

 

 

 

0 ----- . . _ I l
‘_ 2.373‘ '— "'-_. —
v(7) _ v(7) ....................... _
3
.— 2.3127410 _, l I
1 .3 7 6

5. Comparison of Social Welfares
Deﬁne N s Wil —Wf. Then, by using equations (3.25) and (3.44), we can simplify

N as the following

_ (A —a)2(372 -67+4)(—2172 + 227—4)
2(27-0’197—8)’

 

N (A.4)

A negative sign of N indicates that the host country’s welfare under foreign direct

investment is greater than under licensing contracts with the incumbent company. The
sign of N depends on (372 —67+4)(—2172 +227—4) . Once again, deﬁne

b(7) a (372 —67 + 4)(—2172 + 227 -4). The values of b(7) are always negative under

155

7>§ since 2g—(Z/lz—25273+57672—45674-112<0 and b(l.3)=—l3.83.

7

Figure 3.9: The Function of b( 7)

 

 

 

 

0 " ' ' - - -T. .. I
.—13.83, __ ___
............ 7' -'-
b(7) """"""" "-_

b(7) _ .2

4
.— 4.773x10 _, I I -,

1.3 'r 6

6. Derivation of h = O

The incumbent company’s share of operating proﬁts under licensing contracts with

the multinational ﬁrm is represented by

_ 2A_ 2
(l—6)(A—qil—a+til)qil=(1 227-(8)20!) (A.5)

 

By equating equation (A.5) with equation (3.24), we obtain the 6

4574 —1273 -10572 +887-16
(8172—1447+64)72

 

9(7): (A6)

By using equation (A.6), we can draw Figure 3.3.

7. Comparison of Social Welfares

We deﬁne K 5W] —Wf. Then, K can be obtained from equations (3.62) and

(3.63).

156

K
8(87-5)2(87-9)2

= (A —a)2(87 -15)(—6472 + 2727 - 195)

(A.7)

 

A positive sign of K means that the host country’s social welfare under licensing is

greater than under foreign direct investment. To illustrate more easily, again deﬁne

2(7)
8(87 - 5)2 (87 - 9)2

Figure 3.10: The Comparison of Social Welfares

- 2 _
a (87 15X—647 + 2727 195) . By using this formula, we can draw Figure 3.10.

 

 

 

 

 

_ I
0.003 '1.
0 B C .
3': '1’) .’15/8 3.34'- ._
g. .......................... 4,, 8......
- 13.003 l

 

 

 

1.5

Figure 3.10 shows that if 7 is in range B (%<7<%), then the sign of 2(7) is

negative and thus the host country’s welfare under foreign direct investment is greater

than under licensing. Similarly, the host country prefers foreign direct investment to an

arm’s length arrangement when 7 is in range D (7 > 3.34 ). On the contrary, if 7 is

located in range C (182 < 7 < 3.34 ), then the host country prefers licensing to foreign

direct investment.

157

 

REFERENCES

BlomstrtSm, Magnus, and Ari Kokko (1997) “The Impact of Foreign Investment on Host
Countries: A Review of the Empirical Evidence,” World Bank Policy Research
Working Paper #1745.

Chin, Judith C., and Gene M. Grossman (1990) “Intellectual Property Rights and North-
South Trade,” In: Jones, Ronald W., and Anne 0. Krueger (Eds), The Political
Economy of International Trade: Essays in Honor of Robert E. Baldwin, Blackwell,
Cambridge, MA.

Coase, R. H. (193 7) “The Nature of the Firm,” Economica, 4(16), 386-405.

Contractor, Farok J. (1985) Licensing in International Strategy: A Guide for planning
and Negotiations, Westport, Conn. Quorum Books.

Dasgupta, Partha (1986) “The Theory of Technological Competition,” in: Stiglitz, J. E.,
and G. F. Mathewson (Eds), New Developments in the Analysis of Market
Structure, The MIT Press, Cambridge, MA, 519-547.

Gallini, Nancy T., and Brian D. Wright (1990) “Technology Transfer under Asymmetric
Information,” The Rand Journal of Economics, 21 (l), 147- l 60.

Gilbert, Richard J., and David M. G Newbery (1982) “Preemptive Patenting and the
Persistence of Monopoly,” The American Economic Review, 72 (3), 514-526.

Ginarte, Juan C., and Walter G. Park (1997) “Determinants of Patent Rights: A Cross-
National Study,” Research Policy #26, 283-301.

Hanson, Gordon H. (2001) “Should Countries Promote Foreign Direct Investment?” G-
24 Discussion Paper #9.

Horstrnann, Ignatius J., and James R. Markusen (1987) “Licensing versus Direct
Investment: A Model of Internationalization by the Multinational Enterprise,” The
Canadian Journal of Economics, 20 (3), 464-481 .

 

 

, and (1992) “Endogenous Market Structures in

International Trade,” Journal of International Economics, 32, 109-129.

158

 

Mansﬁeld, Edwin, and Anthony Romeo (1980) “Technology Transfer to Overseas
Subsidiaries by U.S.-Based Firms,” The Quarterly Journal of Economics, 95 (4),
737-750.

Markusen, James R. (1995) “The Boundaries of Multinational Enterprises and the
Theory of International Trade,” Journal of Economic Perspectives, 9(2), 169-189.

 

(2001) “Contracts, Intellectual Property Rights, and Multinational
Investment in Developing Countries,” Journal of International Economics, 53,
189-204.

Maskus, Keith E., Kamal Saggi, and Thitima Puttitanun (2004) “Patent Rights and
International Technology Transfer through Direct Investment and Licensing,” In:
Maskus, Keith E., and Jerome H. Reichman eds., International Public Goods and
the Transfer of Technology under a Globalized Intellectual Property Regime,
Cambridge University Press.

Neary, J. Peter (2002) “Foreign Direct Investment and the Single Market,” The
Manchester School, 70 (3), 291-314.

Norback, Pehr-Johan (2001) “Multinational Firms, Technology and Location,” Journal
of International Economics, 54, 449-469.

Pack, Howard, and Kama] Saggi (1997) “Inﬂows of Foreign Technology and
Indigenous Technological Development,” Review of Development Economics, 1 (1),
81-98.

Petit, Maria-Luisa, and Francesca Sanna-Randaccio (2000) “Endogenous R&D and
Foreign Direct Investment in International Oligopolies,” International Journal of
Industrial Organization, 18, 339-367.

Saggi, Kamal (1999) “Foreign Direct Investment, Licensing, and Incentives for
Innovation,” Review of International Economics, 7 (4), 699-714.

Shapiro, Carl (1985) “Patent Licensing and R&D Rivalry,” The American Economic
Review, 75 (2), 25-30.

Teece, D. J. (1977) “Technology Transfer by Multinational Firms: The Resource Costs
of Transferring Technological Know-How,” The Economic Journal, 87 (346), 242-

159

 

 

261.

Vermeulen, Freek, and Harry Barkema (2001) “Learning through Acquisitions,”
Academy of Management Journal, 44 (3), 457-476.

Vishwasrao, Sharmila (1994) “Intellectual Property Rights and the Mode of Technology
Transfer,” Journal of Development Economics, 44, 381-402.

Yang, Guifang, and Keith E. Maskus (2001) “Intellectual Property Rights, Licensing,
and Innovation in the Endogenous Product-Cycle Model,” Journal of International

Economics, 53, 169-187.

Zigic, Kresimir (1998) “Intellectual Property Rights Violations and Spillovers in North-
South Trade,” European Economic Review, 42, 1779-1799.

160

 

   

1"13111111110171111119131