|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
FOREIGN DIRECT INVESTMENT AND THE INVESTMENT CLIMATE
OF DEVELOPING COUNTRIES IN THE WESTERN HEMISPHERE
ERIC T. BONNETT
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Eric T. Bonnett
This dissertation is dedicated to my lovely wife, Rita.
I would first like to thank the members of my supervisory committee for making
my time as a graduate student at the University of Florida as enjoyable as it has been.
During my years in the Food and Resource Economics Department I became distinctly
aware of how well respected Drs. Timothy Taylor, Gary Fairchild, P.J. van Blokland and
Roy Crum are among those in the academic world. More recently, I have also come to
realize the level of respect those outside of academia possess for these men. Their
contributions to this project are greatly appreciated and will never be forgotten. Also I
would be remiss if I did not mention my appreciation for the hours Dr. Ronald Ward
spent assisting me with this research project.
My parents deserve special recognition for instilling in me the confidence to pursue
such an ambitious goal as writing a doctoral dissertation. While they have always
encouraged me to aggressively pursue my goals, they also taught me the importance of
being able to maintain a sound conscience. I will forever be in debt to my mother and
father for passing these most admirable personal traits along to me.
As I find myself at the terminal point of this project, I am reminded of a very
special individual who was taken from this world in a most untimely manner. If it were
not for Dr. Patrick Byrne, I most likely would have never attended graduate school. He
was a source of inspiration for many students during his tenure at the University of
Florida, and his legacy lives on with each of them.
Finally, I would like to thank my wife, Rita, to whom this dissertation is dedicated.
Just as she is essential to my happiness, her unremitting encouragement and support were
essential to the successful completion of this project. I shall always treasure her
commitment to our relationship, and I will forever carry the knowledge that I am a better
person because of her.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
LIST OF TABLES .................................................... ....... .. .............. viii
LIST OF FIGURES ......... ......................... ...... ........ ............ ix
A B STR A C T ................................................. ..................................... .. x
1 IN TR OD U CTION ............................................... .. ......................... ..
Impacts of FDI on Developing Host Countries..........................................................3
P problem atic Situation ............ .............................................................. ........ .. .. ....
Problem Statem ent .................. ......................................... ................ .8
O b j e c tiv e s ........................................................................................................1 0
S c o p e ..............................................................................1 1
2 H ISTORICAL CON TEX T ................................................... ........................ 13
A Taxonom y of Capital Flow s .......................... ..................... ................... 15
A Method of Characterizing Different Types of Capital Flows.............................18
The Structure and Evolution of Capital Flows from 1970 to 2001 ............................22
The Andean Community ...................... ............................... 24
The Central American Common Market..........................................................28
The Caribbean Community and Common Market............................................32
M ER CO SU R .......... .. ........................... ............ .. ........ .... 35
M exico and Chile ........................................ .. .. .... ........ ..... .... 37
Sum m ary of Trends .......................... .............. ................. .... ....... 40
3 THEORETICAL FOUNDATIONS ........................................ ....................... 42
Neoclassical Economic Theory vs. Strategic Management Theory .........................44
D running's O LI Fram ew ork ............................................... ............................. 48
4 EM PIRICAL CON SIDERATIONS ........................................ ....................... 56
The MNC's Motivations for Making Direct Investments .......................................56
Decisions on the M ode of Organization............... .............................................60
Decisions on Location ......................................... ..........................63
The Em pirical M odel .................. ..................................... ...... .. .............. 68
The Host-Country Investm ent Climate .................................... ............... 69
The Policy and Regulatory Environm ent .................................... .................70
P political risk......................... ......... .................................. 73
Factors that affect the return on investment.............................................. 75
C control V ariab les.......... ...................................................................... ........ .. ....... .. 7 8
5 MODEL SPECIFICATION, ESTIMATION, AND EMPIRICAL ANALYSIS .......79
D ata and Sources ...................................................... 79
M odel Specification and Estim ation....................................... ......... ............... 81
R results of E stim action .......... .... .. ........................ .......... ...... .. ....... .... 85
M ERCOSUR ............ ................................ ....................90
The Andean Community .................................. .....................................96
C A R IC O M ......................... ...................... ................... ......... 101
The Central American Common Market........... .................................107
M exico and Chile ...... ........... .................... ......... ................ 113
6 SUMMARY AND CONCLUSIONS..... ........... ......... ...............117
The Least Responsive Countries ................................ .....................117
The Most Responsive Countries................................ ....................119
Policy-R elated C onclusions............................................. ............................. 120
Considerations for Future Research...... ................. ...............122
APPENDIX DEFINITIONS OF SELECTED VARIABLES .......................................124
L IST O F R E FE R E N C E S ......... .. ............. .............................................................. 126
BIOGRAPHICAL SKETCH .............. ............ ............... 133
LIST OF TABLES
4-1 Survey Results on the Importance of Investment Climate Factors ........................70
5-1 Descriptive Statistics by Country (1984-2001)............... .................... ..................80
5-2 Random Effects Regression Results ............................................. ............... 86
5-3 Investment Climate Variables MERCOSUR ............. ..... ..................90
5-4 FDI Responsiveness MERCOSUR ...... ................ ...............90
5-5 Investment Climate Variables Andean Community......................................96
5-6 FDI Responsiveness Andean Community............... ..........................................96
5-7 FDI Responsiveness CARICOM ................................................ 102
5-8 Investment Climate Variables CARICOM ............................... ............... .103
5-9 Investment Climate Variables CACM ..................................... .................108
5-10 FD I Responsiveness CA CM ........................................ .......................... 108
5-11 Investment Climate Variables Mexico and Chile ......................... ............113
5-12 FDI Responsiveness Mexico and Chile ....... ....................................114
LIST OF FIGURES
2-1 Williamson's Key Characteristics of Capital Flows .............................................21
2-2 Capital Flows to Developing Economies in the Western Hemisphere ....................23
2-3 Private Capital Flows to Developing Countries in the Western Hemisphere ..........23
2-4 Capital Flows to the Andean Community ......................................... ...........25
2-5 ODA Flows to Members of the Andean Community .............................................26
2-6 Private Capital Flows to the Andean Community................. .....................26
2-7 Capital Flow s to CA CM ................................................. .............................. 29
2-8 ODA Flows to M embers of CACM ........................................ ...... ............... 30
2-9 FD I Flow s to M em bers of CA CM ..................... ................................................ 31
2-10 Capital Flows to CARICOM ............ ................................. ................. 34
2-11 Private Capital Flows to CARICOM ................................................. 34
2-12 Capital Flow s to M ERCO SU R ........................................ .......................... 36
2-13 Private Capital Flows to MERCOSUR ......................................... ...............37
2-14 C capital F low s to M exico ............................................................... .....................38
2-15 Private Capital Flow s to M exico.................................... ........................... ......... 39
2-16 Private Capital Flow s to Chile ........................................... .......................... 40
3-1 M odes of Internationalization ............................................ .... .......... 51
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
FOREIGN DIRECT INVESTMENT AND THE INVESTMENT CLIMATE OF
DEVELOPING COUNTRIES IN THE WESTERN HEMISPHERE
Eric T. Bonnett
Chair: Timothy G. Taylor
Cochair: Gary F. Fairchild
Major Department: Food and Resource Economics
The structure of capital flows to developing countries around the globe has changed
considerably since the 1970s. In the Western Hemisphere specifically, foreign direct
investment (FDI) accounted for a much larger share of total inflows than any other type
of capital by the end of the 1990s. This trend has not escaped recognition by economic
theorists, and a large literature has emerged as researchers attempt to understand why
some countries attract more FDI than others.
This study examines the impact of changes in the investment climate on a country's
ability to attract FDI flows relative to other countries. Specific attention is given to the
developing countries of the Western Hemisphere. The investment climate is broadly
defined as (i) governmental policies and regulations that affect the relative "openness" of
the country to FDI, (ii) factors that impact the potential return on capital to foreign
investors, and (iii) the level of political risk and corruption in the host country.
Analyzing the impact of changes in these types of factors over time revealed some
interesting results. First, there is evidence to suggest that the state of the investment
climate is an important consideration for foreign investors. Second, FDI in large (in
terms of GDP), relatively unstable economies tends to be the most responsive to small
changes in the investment climate. Conversely, in small economies that either receive
substantial amounts of official development assistance or are dominated by a single
industry (e.g., the production of oil), FDI tends to be less responsive to changes in the
Finally, the results make clear that, across the sample of countries, there is no
single model that can explain all of the differences in the level of response to changes in
the investment climate. In other words, the relative impact of investment climate
variables on the level of FDI differs according to the economic, political and social
conditions inherent to each country. Thus, it is likely that future research into the issue
will yield more interesting results if analysis is conducted on a country-by-country basis.
The volatility of capital flows to developing countries in Latin America and the
Caribbean is well established. Investors in the region often hold short-term assets and
withdraw their funds at the first sign of economic distress (Rojas-Suarez and Weisbrod
1996). This, in turn, contributes to the vulnerability of Latin American and Caribbean
economies to external economic shocks such as changes in commodity prices and
international interest rates (Birdsall and Lozada 1996).
The financial crises that struck many developing countries after the surge in capital
flows during the 1990s led to increased skepticism about the benefits of attracting foreign
capital (World Bank 2001). However, the literature makes clear that just as there are
many different ways to define a financial crisis, there are equally as many theories of
what causes them. Nevertheless, it is generally agreed that the sharp rise in incidence of
financial crises in the last 20 years is not independent of the observed increase in the
magnitude and frequency of international capital flows.
From 1970 to 1992, Latin American and Caribbean economies were between two
and three times as volatile as industrialized economies (Hausmann and Gavin 1996).
While much of the observed volatility is a product of inconsistent macroeconomic policy,
other factors also had an influence. According to Hausmann and Gavin (1996):
another reason for the volatility of Latin American [and indeed, Caribbean]
macroeconomic outcomes is the large external disturbances that routinely buffet the
region. The most important of these are sudden changes in the terms of trade and
in international capital flows. (Hausmann and Gavin 1996, p. 27)
Large capital inflows cause rapid monetary expansion, inflationary pressures, real
exchange rate appreciation and widening current account deficits (Calvo et al. 1996).
Consequently, when flows are interrupted, the current account and exchange rate
experience reverse adjustments (Hoti 2002). Furthermore, these terms-of-trade shocks
have large and statistically significant effects on the variance of economic growth rates
over time (Easterly et al. 1993).
In many of the developing economies in the Western Hemisphere, FDI has become
the dominant component of capital inflows, accounting for a much higher percentage of
gross domestic product (GDP) in the 1990s than in the previous two decades. However,
there are exceptions to this rule, and it is often those countries that are the least
economically and politically stable that are the least successful at attracting FDI. Some
have argued that economic instability leads to uncertainty for the firms and investors who
make direct investments. As a result, investors choose to locate in less "risky" countries.
The cross-country variation in FDI since the early 1990s (discussed further in
Chapter 2) serves as the inspiration for this study. While a large body of empirical
literature addresses the global factors that cause expansions and contractions in global
FDI flows, less attention has been given to explaining the differences in FDI flows across
countries. More specifically, there is a need for additional insight into the factors that
affect the distribution of FDI flows across the developing countries of the Western
Empirical research suggests that the package of assets that accompanies FDI brings
many benefits to the host country including economic growth, development of domestic
industries, increased employment and a higher standard of living (Parry 1973).
Theoretical support for each of these contentions has developed over time while there
have been fewer attempts to argue against the benefits of FDI. What is made clear by
attempts to test the theory is that the benefits realized by the hosts of FDI funds are
conditional upon many factors.
Impacts of FDI on Developing Host Countries
The increase in foreign equity flows to developing countries in the Western
Hemisphere reflects the fact that countries have increased the number of investment
opportunities by promoting the privatization of government enterprises and facilitating
the development of deeper and more liquid financial markets (Moreno 2000). However,
Wells, Jr. (1998) points out that although tensions between foreign investors and
developing host countries had clearly weakened by the mid 1990s, there were few
definite conclusions as to the net impact of FDI in the lesser developed countries (LDCs).
Since then, a vast literature on the subject has emerged.
Much of the research into the economic impacts of FDI is aimed at examining the
indirect or "spillover" effects of hosting direct investment. It has been argued that
indirect effects primarily take the form of technological spillovers and competition
effects. In fact, Blomstrom and Kokko (2003) contend that the expectation that
technological know-how will spillover from multinational corporations (MNCs) to
indigenous firms has been strong enough for host countries to lower barriers to entry,
open up new sectors to foreign investment and, in some cases, provide investment
incentives to foreign firms. Nonetheless, empirical research into the matter is fraught
with conflicting results.
For instance, in a study on manufacturing industries in Mexico, Blomstrom (1986)
found that the presence of foreign MNCs in an industry was positively correlated with
structural efficiency. However, the results did not suggest that the increase in efficiency
was due to the transfer of technology from the MNCs to domestic firms (whether through
imitation or labor migration). Rather, increases in competitive pressure appeared to have
a more important effect.
Aitken and Harrison (1999) found that "foreign equity participation [in Venezuelan
plants] is positively correlated with plant productivity [but,] foreign investment
negatively affects the productivity of domestically owned plants" (p.605). The authors
indicate that in the presence of these offsetting effects, the net impact of FDI on the
productive efficiency of the host economy as a whole is negligible. More recent research
suggests that the extent to which foreign participation impacts the productivity of
indigenous firms appears to depend upon the type of linkages that are examined.
While Aitken and Harrison (1999) examine the effect of FDI on the productivity of
all firms in the economy, Smarzynska (2002) limits her analysis to the effect on upstream
firms (i.e., the backward linkage between foreign affiliates and their local suppliers).
Pointing out that spillovers "are more likely to be vertical rather than horizontal in
nature1" (Smarzynska 2002, p.2), she states, "a rise of ten percent in the foreign presence
in downstream industries is associated with a 0.38 percent increase in output of each
domestic firm in the upstream sector" (Smarzynska 2002, p.16). Furthermore,
Smarzynska found no difference between the effect of wholly-owned foreign subsidiaries
and joint ventures with both foreign and domestic investors.
1 This is due to the multinational firm's incentive to transfer knowledge to its suppliers in order to increase
the quality or decrease the price of its inputs, and prevent the horizontal leakage of information that might
enhance the performance of its competitors.
It has been shown by some researchers that controlling for certain domestic
conditions leads to more conclusive findings on the net impact ofFDI. One such case is
that of the relationship between FDI and economic growth in developing host countries.
Although empirical findings on the nature of this relationship do vary, the World Bank
(2001) indicates that differences in host countries' absorptive capacity2 accounts for a
majority of the variance. In fact, where absorptive capacity is high, it is generally agreed
that FDI has a positive impact on productivity and hence economic growth (World Bank
The study by Borenzstein et al. (1998) serves to illustrate this point more clearly.
Analyzing data on FDI flows to 69 developing countries over a period of 20 years, the
authors found that FDI had a positive effect on economic growth. Furthermore, the
magnitude of this effect was shown to depend strongly upon the available stock of human
capital, a variable commonly used as a measure of absorptive capacity. That is, higher
levels of human capital caused FDI to have a larger positive impact on growth.
Furthermore, with sufficient levels of human capital, FDI appeared to be more productive
than, and complementary to domestic investment. The authors also note that the positive
impact of FDI persisted even after controlling for initial income, human capital,
government consumption and the parallel market premium for foreign exchange.
Aside from the results of a handful of empirical studies, it is generally agreed that
FDI has the potential to facilitate the transfer of ideas from industrialized to developing
countries, thus increasing productivity in the latter (World Bank 2001). In addition to the
stability of FDI relative to other forms of foreign capital, the potential for increased
2 Proxies for absorptive capacity include but are not limited to openness to trade, the amount and quality of
infrastructure and human capital, and inflation.
productivity provides some rationale for why many developing countries have shown an
appetite for FDI.
It has been suggested by the Economic Commission for Latin America and the
Caribbean (ECLAC 2002) that the high volatility of net capital inflows (other than FDI)
during the last ten years has undermined the stability of economic growth in the region.
When net capital inflows are high, domestic credit and liquidity grow too quickly, while
the opposite is true in times of recession (ECLAC 2002). This phenomenon tends to
amplify boom-bust cycles, thus destabilizing economic growth. As an example, Mishkin
The financial crises that struck Mexico in 1994 and the East Asian countries in
1997 led to a fall in the growth rate of GDP on the order often percentage points.
The financial crises in Russia in 1998 and Ecuador in 1999 have had similar
negative effects on real output. Not only did these crises lead to sharp increases in
poverty, but to political instability as well. (Mishkin 2001, p.1)
The link between capital flow volatility and economic growth has led many
researchers to gain interest in the factors that drive capital flows. There is considerable
debate over whether capital flows to developing countries are driven by forces external to
the countries themselves, or by domestic factors.3 The existing empirical literature
suggests that the answer depends upon how the issue is addressed and that analyzing the
sum of all capital flows may be misleading. After all, each form of capital has its own
characteristics and set of circumstances (Lusinyan 2002).
Foreign direct investment accounts for a much larger share of GDP than any other
form of foreign capital inflow for many of the developing countries in the Western
3 Moreno (2000) provides one of the more concise reviews of the empirical literature addressing this issue.
Hemisphere, and much of this investment activity has been driven, or at least encouraged,
by the actions of host-country governments. In addition to privatizing government
enterprises, developing countries have adopted policies intended to foster direct
investment by MNCs. It is therefore surprising that the empirical literature does not
contain more models that attempt to measure the effect of these policies.
Gabel and Bruner (2003) report that multinational corporations and their
subsidiaries employ nearly 200 million people, generate $1.5 trillion in wages and pay
over $1.2 trillion in taxes to their host governments, annually. The almost overwhelming
global presence of the MNC has inspired considerable debate among the many players
involved in international commerce. Government policymakers are interested in
maximizing the benefits of hosting multinationals while simultaneously limiting their
negative impacts. Multinational managers are interested in maximizing the profitability
of their businesses given a set of available investment alternatives. In a world where
MNCs are increasingly prevalent, economists are left with the task of analyzing their
causes and impacts. The resulting body of academic literature is substantial and diverse
Wells, Jr. (1998) indicates that the "issues that face managers public and private -
who are concerned with FDI and the developing countries .. have not attracted
sufficient attention from economists" (p.101). Foreign direct investment flows are
ultimately driven by multinational corporations, and hence, their managers. Thus, the
interaction between government policy and the strategic allocation of capital by MNCs
needs to be addressed. Much of the thought on this matter has been aimed at modeling
the effect of exchange rate regimes on the international allocation of production, but the
issue is much more broad.
The study of foreign direct investment has attracted the attention of researchers
from various fields of business. Economists have portrayed FDI as the result of
production cost differentials and comparative advantages in resource supply. In the field
of finance, FDI has been described as a tool for international portfolio diversification.
Finally, in the strategic management body of literature, FDI patterns are seen as being
dictated by the organizational decisions of the firm's management. Although these
approaches may differ from one another, efforts to understand why FDI occurs have
yielded some compelling results in each line of research.
Foreign direct investment can be viewed as the allocation of financial capital from
an entity in one country to establish or support a business entity in another country.
When defined in this manner, FDI becomes the result of a strategic decision guided by
the goals of management.4 To the extent that foreign investment policy can create an
environment that either fosters or impedes direct investment, the actions of host-country
governments may have a significant influence on FDI flows. Along these lines, the
ultimate concern of this study determining the extent to which FDI flows are affected by
the investment climates of developing host countries.
By surveying a group of U.S.-based multinational companies, Basi (1966)
identified several variables related to the investment climate that were of significant
importance to the investment decision. Of the variables considered by Basi, the
4 "Management" is defined broadly as those individuals who have a direct influence over the international
allocation of direct investors' capital.
following factors were most often cited as either "crucially important" or "fairly
important" to multinational managers: (i) the host country's attitude toward foreign
investment; (ii) political stability; (iii) limitations on ownership; (iv) currency exchange
regulations; (v) the stability of foreign exchange; and, (vi) tax structure. While some of
these variables are difficult to quantify, Basi at least provides an illustration of the
"investment climate" that is consistent with the how the term is used in this study.
The contribution this study makes to the existing literature centers on two primary
characteristics of the project's design. First, the empirical model presented here
examines the relative effect of five distinct investment climate factors on FDI. While
some proxy for the investment climate is often included as an explanatory variable in
existing models, fewer models have been designed to simultaneously examine different
aspects of the investment climate. The second distinguishing characteristic of the study is
its focus on Latin America and the Caribbean. Much of the existing literature is devoted
to examining FDI flows to the entire universe of developing countries. However, less
ambiguous conclusions may be provided to policymakers in the region by limiting the
focus of the analysis to the developing countries of the Western Hemisphere.
There are three reasons for examining the effect of the overall investment climate
on inflows of FDI. First, early attempts to survey firms on the determinants of
international production show that variables related to the investment climate are among
the most influential factors from the perspective of management.5 Second, debate on the
importance of many investment climate factors to foreign investors is ongoing. Finally,
SSee Dunning (1973) for a review of surveys conducted by various authors prior to 1973.
the relevance of the issue to policy-makers is clear, as many investment-climate variables
are under the direct control of the regulatory bodies of host countries.
Examination of FDI flows to the developing countries of the Western Hemisphere
is particularly relevant. The financial markets of these countries are increasingly
accessible to foreign investors, especially those who reside in the United States, where
nearly 20% of world FDI outflows originated during the period 1970-2002. Furthermore,
the provisions of the Free Trade Agreement of the Americas (FTAA), which is scheduled
for entry into force by the end of 2005, provide for the promotion and protection of
foreign investment funds, regardless of their source.
The general objectives of this study are
* To provide an understanding of FDI as an economic phenomenon resulting from
the strategic activities of multinational corporations
* To examine the effect of the investment climate of developing countries in the
Western Hemisphere on inflows of FDI over the last two decades.
In order to achieve these rather broad goals, the study is designed with several
specific objectives in mind. The specific objectives of the study are
* To examine the evolution of the multinational corporation (MNC) from a
theoretical perspective and address the strengths and weaknesses of several streams
* To develop a framework for understanding FDI that integrates the elements of
economic and strategic management theory
* To develop a policy-relevant model of FDI inflows by including variables that are
not only considered as elements of the investment climate, but are also affected by
the actions of host-country policymakers.
Chapter 2 establishes the context for the study by providing an historical account
of contemporary developments in global economic integration. Specific attention is
given to the developing countries in the Western Hemisphere. The chapter compares and
contrasts the different types of international capital flows and presents a detailed analysis
of the structure of capital flows to Latin America and the Caribbean over the last thirty
Chapter 3 presents the theoretical foundations of the study. The chapter begins
with a short review of the existing empirical research on FDI, although the results and
implications of the models are not the main point of concern. Rather, the models are used
to highlight the distinction between neoclassical economic theory and strategic
management theory, as well as to illustrate the fact that research on FDI has been
conducted in a number of contexts. Specific attention is given to the conceptual
evolution of the firm in each of the two streams of theory. Finally, a review of the most
relevant theoretical literature serves to define the perspective from which the empirical
considerations of the study derive.
Chapter 4 revisits the empirical literature is revisited in more depth. The chapter
reviews a range of models that examine issues related to the international allocation of
economic activity. Some of the models presented explicitly address the economic
determinants of FDI, while others analyze factors that affect the strategic decision-
making processes of multinational corporations. Empirical research on investment
location and the alternative modes of foreign ownership is also reviewed.
Chapter 5 presents the empirical model used to analyze a set of panel data on 21
developing countries in the Western Hemisphere over a period of 18 years. The chapter
compares the results from several alternative estimation procedures and provides a
discussion on the responsiveness of FDI to changes in the investment climates of
different countries. Finally, Chapter 6 summarizes the empirical results and presents a
set of conclusions based on the findings. Chapter 6 also provides suggestions for future
It is difficult to find a piece of recent literature in the field of international
economics within which the term "globalization" does not appear. According to the
Economic Commission for Latin America and the Caribbean (ECLAC 2002),
globalization refers to the denationalization of political, legal and cultural systems, as
well as economic markets. The primary entities driving this process are governments,
private investors and financial institutions (Schmukler 2004), and its socio-cultural
impacts on developing countries are far-reaching, and often difficult to measure. In terms
of economic impacts, globalization has led to reductions in trade barriers between
countries, increased exchange of information and technology, and in many cases, greater
vulnerability to worldwide economic conditions. It should be noted, however, that the
path toward globalization has been anything but uniform across countries.
ECLAC distinguishes between the three traditionally recognized phases of
globalization by comparing the relative levels of capital and labor mobility, the
(non)existence of free trade and international institutions for economic cooperation, and
the extent of standardization among national development models. The first phase of
globalization has its roots in the transportation revolution and began to take shape in the
last thirty years of the 19th century. Transportation costs fell dramatically during this
period, effectively reducing the distance between countries and contributing to increased
mobility of goods and labor (Philippe 2001).
With the Bretton Woods conference in 1944 came the initial movement toward
establishing international organizations for economic cooperation (specifically, the
International Monetary Fund and the World Bank). However, during this (the second)
phase of globalization, disparity remained among national models of economic
organization, and capital and labor mobility were limited. Although the Bretton Woods
agreement was intended to foster the flow of capital across international boarders, it
wasn't until the first oil crisis in the early 1970s that international capital mobility really
expanded (Phillipe 2001). Nevertheless, with the foundation of global institutions for
economic cooperation, the stage was set for an expansion in the trade of goods and
Since 1973, the drive towards international cooperation has affected a real
expansion in international trade of manufactured goods, services and capital. The
increased global presence of multinational corporations also contributed to this
expansion. Most recently, the birth of the information age brought about unprecedented
access to information and communication technologies, further facilitating international
While this story characterizes the path to globalization among industrialized
economies throughout the world, it does not necessarily account for the individual
experiences of many developing economies in the Western Hemisphere. There are four
major economic integration groups in Latin America and the Caribbean: the Andean
Community; the Central American Common Market (CACM); the Caribbean
Community and Common Market (CARICOM); and the Southern Cone Common Market
(MERCOSUR). Mexico, Chile and the Dominican Republic do not hold membership in
any of these four integration groups but do relationships with them. Given that the
Dominican Republic is currently an observer to CARICOM and that a bilateral free trade
agreement exists between the two, the Dominican Republic is grouped with CARICOM
in the ensuing discussion. Mexico and Chile are discussed independently.
A Taxonomy of Capital Flows
According to Liberatori (2003), "international economic integration is per se the
result of both direct and indirect mobility of resources across national borders .
including migration of workers, international trade in goods and services, capital flows
and international production and investment" (p.2). In addition, international trade and
investment flows are the result of both equity and non-equity transactions. In addition to
the distinction between equity and non-equity, capital flows are categorized according to
the source of funds and the conditions upon which they are disbursed (i.e., concessional
vs. interest bearing). The Development Assistance Committee (DAC 2000)1 states that
Official transactions are those undertaken by central, state or local government
agencies at their own risk and responsibility, regardless of whether these agencies
have first borrowed the necessary funds from the private sector. Private
transactions are those undertaken by firms and individuals resident in the reporting
country. (DAC 2000, p.6)
Official capital consists of official development assistance (ODA) and other official
flows (OOF). According to the guidelines set up by the DAC, ODA includes flows to
countries on Part I of the Development Assistance Committee List (which includes all of
the countries in this analysis) that satisfy three criteria: (i) the funds must be provided by
official agencies including state and local governments or their executive agencies; (ii)
1 The DAC is the principal body through which the Organization for Economic Cooperation and
Development (OECD) deals with issues related to cooperation with developing countries. The DAC
statistical reporting directives provide a basis upon which all donor countries should report disbursements
of official capital to developing countries.
the transaction must be administered with the promotion of economic development and
welfare of developing countries as its main objective; and (iii) the funds must be
concessional in nature and convey a grant element of at least 25%. Official development
assistance flows are unique in that the debt service includes no interest payment and the
grant portion of these flows is free of repayment obligation. Other official flows, as
defined by the DAC, include official sector transactions with aid recipients that do not
satisfy the ODA criteria.
Private capital flows accounted for a significant share of resource flows to
developing economies in the Western Hemisphere during the last 30 years. Private
capital includes commercial-bank and trade-related lending, foreign direct investment and
portfolio investment. Bank and trade-related lending is comprised of commercial-bank
lending and other credits extended by foreign lenders in the private sector. The criteria
established by the World Bank dictates that lending by commercial banks that are
wholly- or partly-publicly owned be excluded from bank and trade-related lending.
Portfolio capital flows include both equity and bond investment. Portfolio equity
flows are measured as the sum of country funds, depository receipts, and direct purchases
of shares by foreign investors who own less than 10% of the voting stock of a firm.
Portfolio bond investment consists of any bond issues purchased by foreign investors
who own less than 10% of the voting stock of the issuer.
Foreign direct investment has historically been categorized as a private capital flow
despite the fact that direct investment, as defined by the DAC, may originate from private
or official sources. Another characteristic of FDI is that it takes the form of both equity
and non-equity investment. The World Bank's comprehensive guide to FDI released in
2003 describes some of the ambiguities that have historically existed in the measurement
and classification of FDI flows and establishes a clearer definition of FDI than had
previously been provided. The following discussion draws heavily upon the guide.
According to the World Bank,
direct investors may be individuals; incorporated or unincorporated private or
public enterprises; associated groups of individuals or enterprises; governments or
government agencies; or estates, trusts, or other organizations that own direct
investment enterprises in economies other than those in which the direct investors
reside. (Liberatori 2003, p.3)
Furthermore, the IMF indicates that in order for funds to be classified as direct
investment, "it is necessary to establish a (or show a pre-existing) specific relationship
between the parties involved in the transaction" (Liberatori 2003, p.2), where the term
"specific relationship" refers to the investor's possession or acquisition of a lasting
interest or an effective voice in the management of a direct investment enterprise.
In order to minimize the subjectivity of this definition, a specific empirical
threshold has conventionally been used to separate FDI from portfolio flows. The
threshold defined by the IMF dictates that only those entities in which the foreign
investor has acquired at least 10% of the ordinary shares or voting power qualify as direct
investment enterprises. Foreign direct investments are made through greenfield
investment (e.g., creation of new production capacity), joint ventures, or through mergers
and acquisitions (e.g., the privatization of an existing government-owned enterprise).
The types of capital covered by direct investment include: (i) equity capital; (ii)
reinvested earnings; and (iii) inter-company loans.
All flows of capital reported in this study are "net inflows." This is important
because there is a clear distinction between "net inflows" and "net flows." For instance,
the term "net foreign direct investment flows" refers to nonresident direct investment in
the host economy net of resident direct investment abroad. This is not what this study
examines. Rather, "net inflows of foreign direct investment" (with which this study is
concerned) refers to nonresident direct investment in the host economy net of foreign
direct investment funds withdrawn from the host economy by foreign investors. When
defined in this manner, a negative value for "net inflows of FDI" in any given year would
indicate that repatriation of FDI capital (which is an outflow) exceeded new inflows of
FDI in the host economy during the period. Similarly, a negative value for "net inflows
of bank and trade-related lending" indicates that principal and interest (P&I) payments to
foreign lenders exceeded new loan disbursements to the borrowing country.
Furthermore, a negative value for "portfolio equity" would indicate a certain level of
portfolio capital flight, not that more portfolio capital was invested by resident entities
than received by the host economy.
A Method of Characterizing Different Types of Capital Flows
In addition to the definitional characteristics provided by the Organization for
Economic Cooperation and Development (OECD), capital flows differ in terms of their
implications for both the host country and the investor. In fact, Williamson (2000) states
that there are five characteristics relevant to distinguishing among the various forms of
capital flow: (1) cost; (2) conditionality; (3) risk-bearing; (4) access to intellectual
property; and (5) vulnerability to capital flow reversal. The following discussion of these
characteristics draws heavily upon the work presented by Williamson (2000).
Cost. Official capital, given that it often comes in the form of grants or
concessional lending, is traditionally considered to be the cheapest form of capital
available to developing economies. While there is no clear difference between the cost of
commercial-bank lending and portfolio bond investment, it is generally recognized that
the cost of each is less than official capital. Portfolio equity returns tend to be highly
variable, fluctuating with the performance of stock markets. However, given that
markets typically demand an equity risk premium, it is generally accepted that the
required return on portfolio equity investment is, on average, higher than the return on
foreign debt. Meanwhile, foreign direct investment is believed to be the most expensive
form of capital for developing countries. As evidence of this, Williamson points out that
a conservative estimate of the annual return on United States FDI is about 12.4%.
Conditionality. Developing countries are often required by the lending
multilateral development bank (MDB) to establish a clear set of macroeconomic policy
objectives in order to qualify for concessionary loan funds. In addition, development
funds tend to be tied to the design and implementation of specific development projects.
Thus, disbursements of official development assistance (ODA) are considered as highly
conditional. Furthermore, the negotiations that often take place between multinational
firms contemplating investments and host countries (e.g., performance requirements
imposed by the host) result in FDI also being highly conditional, although perhaps not to
the extent of official flows. Other forms of capital are typically free of these types of
Risk-bearing. "Risk-bearing" refers to who reaps the benefits (or foots the bill) in
the case of unexpectedly high (or low) revenues and costs. Here, risk is broken down
into commercial risk, interest risk and exchange rate risk. Virtually all official capital,
with the exception of investments made by the International Finance Corporation, convey
commercial risk to the borrower. The same is true of bank loans and portfolio bond
investment, except in the extreme case of lender bankruptcy. In the case of portfolio
equity investment and FDI, commercial risk is typically assumed by the investor.
Interest risk is also born solely by the investor in the case of equity investment,
whether portfolio or direct. For official flows and loan capital, the placement of interest
risk depends upon the terms of the loan. Fixed interest rates place the burden on the
lender while floating rates convey risk to the borrower. Given that most official loans
carry fixed interest rates, the lender (in most cases a multilateral development bank)
usually bears the risk of changes in the market interest rate. Both bonds and commercial
loans tend to made at floating interest rates, and therefore the burden is typically assumed
by the borrower.
Exchange rate risk, in the case of equity investment, is usually born by the investor.
Given that developing countries are rarely able to borrow large sums of money in their
local currency, exchange rate risk is typically conveyed to the borrower of official funds
and commercial loan capital. However, this risk may be transferred to the lender if
borrowed funds are denominated in the domestic currency.
Intellectual property. Since the pioneering work on FDI by Stephen Hymer
(1976), it has been commonly asserted that direct investment by foreign firms brings with
it the transfer of technological know-how in the form of patents, trademarks and
managerial expertise. As pointed out in Chapter 1, these externalities associated with
FDI are often referred to as spillover effects. The conditions associated with official
capital (e.g., macroeconomic policy requirements) also result in the transfer of knowledge
from MDBs and other official entities to host economies. Each of the other forms of
capital inflow typically does not facilitate the transfer of intellectual property.
Vulnerability to reversal. Although firms do shift working balances into and out
of different currencies in response to changes in macroeconomic prospects, the sum of
these transfers is likely to be small relative to the total of sunk capital invested in a
foreign economy (Williamson 2000). Thus, FDI is traditionally regarded as being
minimally vulnerable to reversal. Official flows also tend to be stable, and have actually
been shown to be counter-cyclical in nature. Private bank and trade-related lending,
while usually grouped together, tend to be very different in terms of vulnerability to
reversal. Short-term loans are more vulnerable to reversal than long-term loans, and
trade-related credits are almost always disbursed on extremely short terms. However, the
fact that trade credits are constantly renewed as new trade transactions need to be
financed makes trade-related lending more stable forms of capital inflow. At the other
extreme is bank lending, which in recent crises (e.g., the East Asian crisis and the Latin
American debt crisis) has proven to be highly vulnerable to reversal.
Cost of Capital
Vulnerability Portfolio High additionall
*I Loran onditionality
Intellectual Property Risk
Figure 2-1. Williamson's Key Characteristics of Capital Flows
Figure 2-1 summarizes Williamson's framework for distinguishing between
different types of capital flows. The origin (or center) of the diagram corresponds to low
levels of conditionality, risk bearing, etc., while the outer limits represent the highest
levels. The figure shows reiterates the fact that FDI tends to be conditional and, at least
in theory, facilitates the transfer of technology. In contrast, portfolio equity investment
brings with it no conditionality or transfer of knowledge. The cost of equity investment
is typically high, which reflects the investor's burden of commercial risk, interest rate
risk and exchange rate risk. Perhaps most importantly, FDI is considered the most stable
form of foreign capital inflow available to developing host countries.
The Structure and Evolution of Capital Flows from 1970 to 20012
Figure 2-2 shows the structure of capital inflows to most of the developing
economies of the Western Hemisphere during the period 1970-2001. What is readily
evident in the figure is that private investment and lending served as the major sources of
foreign financing to these countries for much of the period 1970-2000. With the
exception of the mid-1980s, inflows of private funds were generally 3- to 5-times greater
than official development assistance and other official flows combined.
Looking specifically at the structure of private flows, bank and trade-related
lending was the most abundant type of private finance throughout the 1970s and early
1980s. However, this trend came to an abrupt end as the Latin American debt crisis set
in. As Figure 2-3 shows, repayments of private foreign debt actually exceeded new
lending in 1989 and 1993. The global economic slowdown of the late 1990s also had a
detrimental effect on the level of private lending to the developing countries of the
2 All dollar-denominated figures are stated in terms of current US dollars. These figures are converted from
domestic currencies using single-year official exchange rates.
Western Hemisphere. Similarly, net inflows of portfolio capital were rather unstable,
displaying the greatest year-to-year volatility during the 1990s.
20 Capital Flows
o Other Official
= Flow s
0 [: -- -- -
Figure 2-2. Capital Flows to Developing Economies in the Western Hemisphere
5/ ---- Foreign
I \ Bank and
S \V Lending
o 15 -
m _. Investment
Figure 2-3. Private Capital Flows to Developing Countries in the Western Hemisphere
In contrast to bank lending and portfolio investment, inflows of FDI grew during
the period 1970-2001; a majority of the growth occurred in the period immediately
following the debt crisis. Inflows of FDI stood at 0.99% of GDP in 1993, while they
accounted for just under 5% in 1999. However, the most remarkable trend is that inflows
of FDI, which once represented less than 20% of the size of private lending, ended the
period at over 50-times the magnitude of private lending and portfolio investment
Overall, the information presented in Figures 2-2 and 2-3 provides a reasonably
good characterization of how the structure of capital inflows to developing countries
across the globe has evolved over the last three decades. However, some of the most
interesting inferences are revealed by comparing the cases of individual countries.
Accordingly, the remainder of this chapter is devoted to examining the different
experiences of developing countries in the Western Hemisphere.
The Andean Community
The Andean Group was established in 1969 by the Agreement of Cartagena and
originally included Bolivia, Columbia, Ecuador, Peru and Chile (Andean Community
General Secretariat 2003). Venezuela became a party to the Agreement in 1973 and
Chile subsequently withdrew in 1977. The Group was primarily established as a reaction
to the poor performance of the Latin American Free Trade Agreement, but by the late
1980s, commerce orchestrated by the Agreement amounted to no more than 5% of the
combined trade of the group's members (Hanratty and Meditz 1989).
The Group's adoption of Decision 220 in 1987 loosened foreign investment
regulations, allowing greater freedom for the repatriation of profits, a higher percentage
of foreign ownership and investment in a wider variety of industries (Hanratty and
Meditz). The name of the organization was changed to the Andean Community as pat of
The Trujillo Amending Protocol of 1996. In addition to a common external tariff and
rules to prevent distortions in competition, the present-day agreement includes measures
to guarantee international investments.
As Figure 2-4 shows, capital flows to the Andean Community have, for the most
part, been dominated by private investment since 1970. Official development assistance
flows to the group were less than 1% of GDP in every year except 1991, and even then
they only reached 1.10% of aggregate output. Similarly, other official flows (OOF)
exceeded 1% of GDP only twice in 1990 and 1992 and never reached higher than
1.32% of GDP. Overall, both ODA and OOF (as a percentage of GDP) were generally
flat between 1970 and 2001 and never accounted for more than 0.5% of GDP.
Capital Flow s
S.- Other Official
0% 7' Flow s
Figure 2-4. Capital Flows to the Andean Community
Figure 2-5 shows that ODA flows (as a percentage of GDP) to Bolivia grew from
1970 to 1992. In fact, ODA to Bolivia peaked at just under 12% of GDP in 1992. In
dollar magnitude, Peru attracted almost as much ODA as Bolivia. However, private
capital flows to the Andean Community as a whole were much more substantial (as well
as more volatile) than official flows. Private investment declined sharply during the first
half of the 1980s, but after the adoption of Decision 220 these flows expanded just as
quickly as they had dropped off. Private lending was the dominant source of capital for
the group until 1985, when foreign direct investment began to grow. The shift in private
capital inflows from debt to equity investment is evident in Figure 2-6.
8% /\ I
8 I \'\
Q p\ \I
.... ... I . ....
Figure 2-5. ODA Flows to Members of the Andean Community
0) 0) 0)0 0) 0
Figure 2-6. Private Capital Flows to the Andean Community
Heavy lending in Venezuela during the late 1970s pushed private lending to the
Andean Community to over 4.70% of the group's aggregate GDP. However, repayments
of foreign debt actually exceeded new lending as Venezuela began to service those
obligations during the 1980s; in 1987, repayments of private debt by the Community as a
whole exceeded new loan disbursements by almost $1.86 billion, leading to a net outflow
of private capital equal to 1.21% of GDP. Net bank lending (i.e., new lending minus
interest and principal payments on existing debt) remained mostly negative until 1993.
Then it was Columbia that attracted a majority of the private lending in the Andean
Community. Although lending to the group in 1997 was actually higher than in 1979,
increased economic activity caused bank and trade-related lending to account for a much
smaller proportion of GDP.
FDI flows to the Andean Community alternated between negative and positive
values from year to year before stabilizing in the 1980s. However, both FDI and
portfolio investment in the Andean Community expanded rapidly following Decision
220. Foreign direct investment surged to over 5% of the group's GDP in 1997 on the
heels of a sharp increase in investment in Columbia and Venezuela. Despite a continuing
surge of FDI in Bolivia, direct investment in the larger countries contracted over the next
few years and FDI in the Community as a whole fell from 5.07% of GDP in 1997 to just
over 3% in 2001.
Organized markets for equity securities existed in the Andean Community as early
as the end of the 16th century when the emergence of joint-stock companies began in
Columbia (Bolsa de Valores de Columbia, 2003). However, foreign portfolio investment
in the Andean Community first occurred in Venezuela in 1989 and equity markets in
Peru, Columbia and Ecuador began to attract foreign portfolio capital in the mid-1990s.
The Andean Community is unique in that portfolio investment in the group during the
last 32 years was split evenly between the equity and bond markets.
In summary, private capital was the most significant and most volatile type of
financial capital flowing into the Andean Community over the past three decades.
Private lending accounted for the majority of private capital flows to the group prior to
the Latin American debt crisis of the 1980s, while FDI and portfolio investment reigned
in the latter part of the 1980s and the 1990s. Disbursements of private debt contracted
sharply in the 1980s and never recovered to pre-crisis levels. Although inflows of
portfolio investment expanded during the 1990s, they paled in comparison to inflows of
FDI and private lending.
The Central American Common Market
The General Treaty of Central American Economic Integration established the
Central American Common Market (CACM) in December 1960. The agreement went
into effect in June 1961 with the completion of the necessary articles of ratification by El
Salvador, Guatemala and Nicaragua. Honduras and Costa Rica subsequently acceded to
the agreement in 1962. The General Treaty established the Central American Bank for
Economic Integration, a common external duty known as the Central American Standard
Import Tariff and a free trade area among CACM members. The Integration Industries
Convention (Regimen de Industrias de Integraci6n-RII) was also established at that
time as a means for governing foreign investment practices. In addition, the Convention
granted special incentives and privileges to firms given "integration industries" status.
Over time, this component of the agreement proved to be the most difficult to implement
The CACM integration process was somewhat successful in the 1960s, but by the
end of the decade Honduras and El Salvador had engaged in the so-called "Soccer War."
This dispute effectively reduced the Central American Common Market to an entity that
existed merely on paper (Merrill 1994). Although it took nearly a decade to establish an
official peace accord between the two countries, Honduras was able to negotiate a set of
favorable trade arrangements with other CACM members. The movement toward the
economic integration of Central America was rejuvenated in the early 1990s and
Honduras officially rejoined the process in early part of 1992.
Figure 2-7 shows the evolution of net capital inflows to CACM over the last three
decades. Bank and trade-related lending kept private capital flows to the group above 2%
of aggregate GDP during most of the 1970s. Lending subsided and ODA flows increased
three-fold throughout the 1980s and into the early 1990s; ODA reached 7.02% of GDP in
6 -- Rivate Capital
a 4% --Flow s
0% l~ /- Other Official
Figure 2-7. Capital Flows to CACM
From that point, inflows of ODA decreased and private capital inflows, led by an
expansion in FDI, jumped from just over 2% of GDP in 1997 to 5.56% in 1998. Inflows
of private capital had subsided by 2001 and were once again exceeded by inflows of
ODA. Other official flows never exceeded 1.5% of aggregate GDP.
As shown in Figure 2-8, the sharp increase in ODA in the early 1990s was led by
foreign aid flows to an ailing Nicaraguan economy. Foreign aid to Nicaragua reached
$837 million in 1991 and accounted for over 50% of the country's GDP. Flows of ODA
fell to 18.2% of Nicaraguan GDP over the next two years, only to rebound to over 48.5%
of GDP in 1996 (or roughly $933 million).
0-- Costa Rica
S....... El Salvador
30 -.-. -Guatamala
10% -- Nicaragua
Figure 2-8. ODA Flows to Members of CACM
Although relatively less substantial than ODA flows to Nicaragua, flows of
assistance to both Honduras and El Salvador were significant during the 1980s and
1990s. As Figure 2-8 shows, flows of ODA to El Salvador reached 10.64% of GDP in
1987 but gradually fell below 2% by 2001. Official development flows to Honduras
reached their highest level in 1999, when they accounted for more than 15% of GDP.
As previously mentioned, private capital flows to CACM were dominated by bank
and trade-related lending during most the of the 1970s. However, as was the case with
most of the countries in the Western Hemisphere, private lending contracted significantly
during the Latin American Debt Crisis. Foreign direct investment in the group, while
fairly consistent throughout the 1970s and 1980s, expanded during the 1990s. In fact,
FDI increased from 1.75% of the group's aggregate GDP in 1997 to over 5% of GDP in
1998. Interestingly, Nicaragua was both the primary recipient of ODA funds and one of
the major beneficiaries of the expansion in FDI.
During the 1990s, major assistance from the International Finance Corporation was
directed at implementing new FDI laws and establishing an investment promotion agency
in Nicaragua. Figure 2-9 shows that FDI flows to Nicaragua increased from 0.0% to
13.23% of GDP during that decade. In 1998, the privatization of electricity and
telecommunications providers in El Salvador contributed to substantial inflows of FDI
(Sandrasagra 2000). Meanwhile, flows to Costa Rica grew steadily and eventually
reached 4.42% of GDP.
10% Costa Rica
S% -- Guatamala
s? 4% ---------------- "---- ^ 'r^ i
4 \ / \ Honduras
2% i00/ ..- ..- .- ., .. ---Nicaragua
Figure 2-9. FDI Flows to Members of CACM
Bond purchases by foreign investors were inconsistent and rarely exceeded 1% of
GDP in a given year among the CACM countries. The most remarkable inflow of
portfolio funds was in 1993, when investment in Honduras rose to 4.37% GDP ($152
million). Flows of portfolio capital to Costa Rica and El Salvador in 2001 led portfolio
investment in CACM to 0.89% of aggregate GDP, the highest level observed in the past
In summary, official development agencies were the main source of financial
capital flows to CACM throughout the 1980s and most of the 1990s. While the group's
ability to obtain private credit was obviously affected by the debt crisis of the 1980s, FDI
as a percentage of GDP increased steadily from 1984 to 1998. It is this expansion in FDI
that brought inflows of private capital to levels that exceeded inflows of ODA in the late
1990s. However, it is noted that the magnitude of private inflows and official inflows
converged during the first two years of the new millennium.
The Caribbean Community and Common Market
The idea of regional integration began in the Caribbean with the establishment of
the rather short-lived British West Indies Federation of 1958. However, at the end of the
four years during which the Federation existed, the 10 members found themselves
economically the same as they had been for centuries (CARICOM Secretariat 2003).
Although plans for a customs union were part of the original Federation, free trade
among the islands in the region was not realized until the Caribbean Free Trade
Association (CARIFTA) came into effect in 1968.
The Caribbean Community Treaty was signed in Chaguaramas, Trinidad on July 4,
1973 and it was agreed among the four independent countries of CARIFTA Barbados,
Guyana, Jamaica, and Trinidad and Tobago that the agreement would come into effect
in August of that year. The revised Treaty of Chaguaramas officially established
CARICOM among these four independent signatory countries. The Treaty also set forth
that eight other Caribbean nations Antigua, Belize (British Honduras), Dominica,
Grenada, St. Lucia, Montserrat, St. Vincent and St. Kitts and Nevis would join the
Community by May 1, 1974. Haiti became the first French speaking Caribbean State to
obtain full membership in CARICOM on July 3, 2002.
With the exception of Montserrat (which remains a British territory), each of the
signatory CARICOM countries is currently independent from European control.
Nevertheless, nearly half of the members achieved independence subsequent to 1975.
The obvious problem with analyzing capital flows to these countries is the lack of
availability, or separability, of such data. As a result, seven of the 15 CARICOM
countries are not included in this analysis. The remaining members do represent an
anecdotal sample of CARICOM, as they account for nearly 75% of the overall economic
activity of the group (as measured by 2001 GDP). It is noted, however, that the unique
experiences of the eight CARICOM countries not included in this analysis may not be
characterized by the experiences of the countries that are examined.
As shown in Figure 2-10, capital flows to CARICOM were dominated by those
from private sources during the first half of the 1970s. However, massive repayments of
private debt obligations by Jamaica in 1976 and 1978 led to net outflows of private
capital in those years. Throughout the 1980s, official development assistance gradually
increased in some of the largest countries in CARICOM (including Haiti, Jamaica, and
the Dominican Republic) and ODA flows to the group reached 3.45% of GDP by 1990.
As time passed, private investment in CARICOM began to strengthen primarily in
Jamaica, Trinidad and Tobago, and the Dominican Republic. By 2001, private capital
flows (as a percentage of GDP) were in excess of the levels seen in 1970. In fact, net
inflows of private capital were almost 10-times the magnitude of net ODA inflows in
\ / L Official
0%~ Flow s
Figure 2-10. Capital Flows to CARICOM
Figure 2-11 shows the structure of private capital flows to CARICOM. Foreign
direct investment was the primary source of private capital for CARICOM as a group
during the period 1970-2001. While bank and trade-related lending never exceeded 1%
of GDP, FDI accounted for more than 7% of GDP in the early 1970s and more than 5%
Figure 2-11. Private Capital Flows to CARICOM
/ ~LC~ ~s
of GDP in the late 1990s. Portfolio investment in CARICOM expanded in both 2000 and
2001, reaching 3.15% of GDP. Most of this investment came in the form of bond
purchases in by foreign investors in Jamaica, Trinidad and Tobago, and Barbados.
MERCOSUR is an acronym for the Southern Cone Common Market and
specifically refers to the South American countries of Argentina, Brazil, Paraguay and
Uruguay. The Treaty of Asunci6n was signed on the March 26, 1991, effectively
establishing MERCOSUR at the end of November that same year. The Treaty was
intended to promote economic integration among countries in the region, although Chile
and Bolivia were conspicuously absent from the agreement (Hudson 1995). Bolivia had
originally intended to become the fifth member of MERCOSUR, although this accession
has never taken place. Furthermore, Chile evaded the agreement on the contention that
the other four signatory countries would have to lower their tariffs to the Chilean level
before Chile would join.
The Protocol of Colonia, which was signed in January of 1994, specifically
addresses the promotion and reciprocal protection of investments among the members of
MERCOSUR. The Protocol grants national treatment to investments in MERCOSUR
signatory countries made by investors from other members of the group. Other protocols
to the Treaty of Asuncion cover the defense of competition, the protection of intellectual
property rights and dispute settlement.
Figure 2-12 shows that for the last three decades financial capital flows to
MERCOSUR were dominated by private flows. Paraguay was the only member of
MERCOSUR for which ODA represented a substantial proportion of GDP, ranging from
0.67% to nearly 3.5%. Although Brazil received more ODA in most years than the other
three members of MERCOSUR combined, the immense size of the Brazilian economy
kept annual receipts to under 0.5% of GDP. Other official flows, while slightly higher
than ODA, never exceeded 0.75% of the group's aggregate GDP in a single year.
5% Capital Flows
1 % Other Official
Figure 2-12. Capital Flows to MERCOSUR
Figure 2-13 shows the evolution of private flows to MERCOSUR. In much the
same manner as the rest of Latin America, private flows to MERCOSUR consisted
mostly of debt throughout the 1970s and early 1980s. The boom in FDI began in 1994
and continued through the end of the decade. Brazil and Argentina attracted a vast
majority of the FDI flows to the group. In fact, FDI in Argentina quadrupled in the latter
half of the 1990s, and FDI in Brazil in 1999 was more than 6-times the level in 1995.
Large flows of portfolio capital also poured into both Argentina and Brazil throughout
the 1990s; flows to these two countries led portfolio investment to just over 3% of
MERCOSUR's aggregate GDP in 1993.
MERCOSUR is unique in that portfolio investment in the group really began to
increase prior to the expansion in FDI, whereas this wasn't necessarily the case in the rest
of Western Hemisphere. With the exception of Mexico, the heaviest portfolio investment
in Latin America took place in Argentina and Brazil. Also, differences in the magnitude
of private capital flows and official flows are also more obvious in MERCOSUR than in
any of the other groups in the Western Hemisphere. However, the gap narrowed slightly
in 2001 as private flows to the group dropped, led mostly by contractions in both
portfolio investment and FDI in Argentina.
5% -- Foreign Direct
e 3% Bank and
0( \ Trade Related
"S 2% --------------------- --------^---- Ledn
7 2/% \---- \ Lending
1% 4 \ \ Portfolio
Figure 2-13. Private Capital Flows to MERCOSUR
Mexico and Chile
Mexico is a party to several free trade agreements with countries and trade groups
on both sides of the globe. The most significant of these agreements are the so-called
Group of Three (1995) and the 1994 North American Free Trade Agreement (NAFTA).
The Group of Three is a sub-regional economic complementarity agreement which
includes provisions for trilateral investment flows between Mexico, Columbia and
Venezuela. Similarly, NAFTA also provides for the national treatment of signatory
countries' investors and their investments.
Figure 2-14 shows that, with the exception of four years in the 1980s, private
funds were the most abundant type of financial capital flowing into Mexico during the
period 1970-2001. In fact, Mexico attracted almost twice as much private capital as the
Andean Community, CACM and CARICOM combined. Private flows to Mexico
consisted mostly of bank and trade-related lending until the mid-1980s. However, credit
reversals and repayments to foreign lenders exceeded new borrowing in 1993 by more
than $6 billion.
6% Capital Flows
a. 4% Official
4 3% IAssistance
2 Other Official
0% ------ -' ----^--^^^
Figure 2-14. Capital Flows to Mexico
Figure 2-15 shows the structure of private capital inflows to Mexico during the
period 1970-2001. The figure shows that portfolio investment poured into Mexico at an
unprecedented rate in 1993, perhaps in anticipation of NAFTA. Bond purchases by
foreign investors amounted to nearly $9 billion while portfolio equity investment soared
to $14.3 billion. Total portfolio investment in 1993 reached 4.87% of GDP, whereas FDI
only amounted to just over 1%. Direct investment expanded in the latter half of the
1990s and eventually reached 4% of GDP in 2001. Meanwhile, portfolio flows were
much more volatile, ranging from -0.38% to 4.29% of GDP.
Chile and Mexico are similar in the fact that ODA accounted for a small portion of
total capital inflows over the last three decades. In the case of Chile, ODA never
surpassed $200 million in any single year. Private capital flows were substantial during
the late 1970s and early 1980s, accounting for anywhere between 9% and 12% of GDP
annually. Nearly all of the private capital flowing into Chile at this time took the form of
bank and trade-related lending. Private credit collapsed during the Latin American debt
crisis and for a period of three years in the late 1980s official flows to Chile actually
exceeded inflows from private sources.
2% A I \ Trade
0 2% Related
0 / I Lending
'""1 / ./ --- Portfolio
0% .--\ \ Investment
Figure 2-15. Private Capital Flows to Mexico
Following the crisis, FDI in Chile surged to 12.3% of GDP. A slight recovery in
private lending and a modest increase in portfolio investment also contributed to private
capital inflows reaching almost 18% of GDP by the end of the 1990s. As Figure 2-16
shows, the new millennium brought a sharp decline in FDI, and thus, net private capital
inflows as a whole.
%-- Foreign Direct
Q /Bank and
Figure 2-16. Private Capital Flows to Chile
Summary of Trends
The figures in this chapter provide a graphical illustration of both the volatility and
evolving structure of capital flows to the developing countries of the Western
Hemisphere. As previously mentioned, Schmukler (2004) argues that the evolutionary
process has been shaped by three primary agents. First, governments have influenced the
structure of capital flows in some instances by relaxing restrictions on the foreign
exchange transactions and allowing increased participation by foreign investors in many
sectors. Second, borrowers and investors, in choosing among different forms of
financing alternatives, have also played a role. Finally, financial institutions have begun
to offer a broader range of financing alternatives to investors and borrowers by making
use of international equity and debt markets. As a result, the developing countries of the
Western Hemisphere by and large have become more financially integrated with the
developed economies of the world. However, some interesting differences have been
exposed in this chapter with regard to the level of private foreign investment in the
various countries within the region.
Private capital flows to developing countries around the globe expanded sharply
during the last thirty years. The bulk of this expansion was accounted for by increases in
foreign direct investment in many Latin American and Caribbean economies. The
chapters that follow examine some of the potential reasons for such substantial increases
in FDI. Chapter 3 is devoted to reviewing the theoretical considerations of researchers
who have examined the issue in the past.
Foreign direct investment has been examined in a number of contexts in the
economic literature. The significance of the topic is illustrated by the variety of
theoretical models and frameworks that have emerged from efforts to characterize and
explain FDI. The diversity of the empirical literature alludes to the existence of many
distinct research agendas. An area that receives a disproportionate amount of attention
(and the area of concern in this study) is identifying the factors that drive flows of FDI.
The economic literature offers several approaches to examining this issue.
Ethier (1986) developed a general equilibrium approach to the determination of
FDI flows, addressing the need to understand how the founding principles of neoclassical
trade theory relate to the issue of foreign investment. Similar models were developed by
Helpman (1984) and Markusen (1984). While they are theoretically elegant, efforts to
test these models empirically have met with limited success. In fact, a review of the
empirical literature on FDI uncovers little in the way of general equilibrium comparative-
A second approach to the determination of global FDI flows is rooted in the theory
of industrial organization and is aimed at explaining why individual firms make
investments in foreign countries to produce the same goods as they produce at home
(Blomstom and Kokko 1997). Early theoretical work in this area, including Hymer
(1976) and Vernon (1966), focused on the firm-specific characteristics that make FDI
more attractive than exporting (Buckely and Casson 1998b). Meanwhile, Dunning (1977,
1995) expanded the set of factors to include host-country-specific variables as well as
firm-specific characteristics. Empirical research along these lines has shown that tariff
and non-tariff barriers to trade, as well as the legal, political and economic conditions of
host countries have a significant effect on investment decisions (Davidson 1980).
In a third body of literature, researchers model FDI flows as a function of either the
global supply of investment funds, the host country's demand for investment funds, or
some combination of the two. In one example of such an analysis, Dasgupta and Ratha
(2000) developed a two-stage approach that addresses both the supply and demand for
investment funds without relying on general equilibrium. In the first stage, the global
supply of FDI is determined by so-called "push" factors1 as investors in developed
countries decide how much capital to invest in the developing countries. In the second
stage, the global supply of direct investment funds is given, while country-specific
variables2 (the so-called "pull" factors) determine each developing country's respective
share of FDI.
The remainder of this chapter is devoted to comparing two widely-cited streams of
theory that are often used to explain the global allocation of FDI flows: neoclassical
economic theory and strategic management theory. This will serve to illustrate the
perspective from which the empirical section of this study is approached. The
methodology developed here is that foreign direct investment, both vertical and
1 Variables considered in the first stage of the process include global market growth (proxied by world
trade as a percentage of world GDP), world GDP growth, GDP growth in developing countries as a whole,
the real LIBOR rate and an index of privatization.
2 Variables examined in the second stage include the current account balance, GNP per capital and private
non-FDI flows as a percentage of GDP.
horizontal, is the observable outcome of coordinated international economic activity3. As
such, the international allocation of FDI flows is ultimately dictated by investors as they
choose among alternative investment locations based on criteria that are perceived to
affect profitability. Neoclassical economic theory pays little attention to the actions of
firms with regard to resource allocation, treating the firm as the proverbial "black box"
into which resources go and out of which goods emerge (Demsetz 1997). Conversely,
the strategic management stream of literature (which has Coase's 1937 theory of
internalization at its core) is concerned with explaining the strategic allocation of
resources by firms, thus providing a role for the decisions of management.
Neoclassical Economic Theory vs. Strategic Management Theory
The purpose of neoclassical economic theory, with perfect competition as its core,
"is to understand price-guided, not management-guided, resource allocation" (Demsetz
1997 p.426). The economic firm is generally embodied in a mathematical function (i.e.,
a production function) defined in terms of technology. In this construct, the usefulness
of the theory lies in its ability to "capture fundamental economic forces and their
interrelationships" (Taylor and Seale 1999 p.8). The theory illustrates a state of
dependency, where production takes place in the firm and consumption and resource
supply take place in households (Demsetz 1997). The strategic management of resource
supply and output demand conditions is not directly addressed by neoclassical economic
theory, giving rise to the suggestion that the economic firm is a "black-box" about which
little is known.
3 Most foreign direct investment is "horizontal" in the sense that a majority of the output of the foreign
affiliate not intended for export to the parent firm's home country (Markusen 1995).
Over time, neoclassical theory has been useful in attempting to explain the
domestic, and by extension in the trade literature, the international allocation of resources
and economic welfare. In neoclassical trade theory, the notion of the firm is generally
that of a purely domestic entity "competing via trade with the national champions of
other countries" (Markusen 1995 p. 169). In addition to the assumption of perfect
competition, extending the theory to the international trade of goods and services often
requires the researcher to assume the existence of comparative advantages in resource
supply. The shortcomings of this extension have been recognized by many and a more
recent research agenda (referred to by Markusen as the "new trade theory") is aimed at
acknowledging that "trade and gains from trade can arise independently of any pattern of
comparative advantage (as traditionally understood) as firms exploit economies of scale
and pursue strategies of product differentiation in an imperfectly competitive
environment" (Markusen 1995 p.169).
Relaxing the assumption of perfect competition does not pose a significant problem
for general-equilibrium trade analysis. However, the definition of the firm in neoclassical
trade theory inherently limits its ability to provide any explanation for the widespread
"real-world" phenomenon of the multinational enterprise. As the boundaries of the firm
expand in both geographic and product space, the need arises to address the strategic
activities that occur inside the black-box of the economic firm.
A useful example is Calderon-Rossell's (1985) attempt to model the effect of
foreign exchange rates and production costs on the MNC's choice between producing in
foreign and/or domestic locations. In this model, the multinational firm is composed of a
parent (in the home country) and a subsidiary (in a foreign country). Monopolistic
demand and cost functions for each are assumed. The limitations of this approach are
characteristic of neoclassical trade model in that the analysis is limited to a two-country
world and requires the assumption of perfect competition, comparative advantage and
zero transportation costs.
The strategic management body of literature was inspired by Robinson's (1932)
pioneering suggestion that assumptions in economic theory should correspond to
conditions in the real world (Coase 1937). Robinson (1932) takes notice of the disparity
between the firm as it is defined in economic theory and the firm as it is described by the
"plain man." Coase (1937) subsequently set out to provide a definition of the firm that is
"not only realistic in that it corresponds to what is meant by a firm in the real world"
(p.386) but is also tractable by the tools of economic analysis.
Coase's (1937) seminal paper addressed the motives for organizing domestic
assets and labor into a firm rather than making use of specialized market exchanges to
conduct arm's-length transactions. The main tenet is that the cost of using the "price
mechanism" associated with market exchanges, along with the costs of negotiating and
concluding separate contracts make it more profitable to internalize production activities.
He stated, "It is true that contracts are not eliminated when there is a firm but they are
greatly reduced" (Coase 1937 pp.390-91).
Coase contended that elements of the regulatory regime in an economy also
provide incentive for firms to internalize operations. Specifically, he contends:
Another factor that should be noted is that exchange transactions on a market and
the same transaction organised within a firm are often treated differently by
Governments or other bodies with regulatory powers. If we consider the operation
of a sales tax, it is clear that it is a tax on market transactions and not on the same
transactions organised within the firm. Now since these are alternative methods of
"organisation" by the price mechanism or by the entrepreneur such a regulation
would bring into existence firms which otherwise would have no raison d'etre...Of
course, to the extent that firms already exist, such a measure as a sales tax would
merely tend to make them larger than they would otherwise be. (Coase 1937 p.393)
Thus, the "Coasian" concept of internalization can be described as the firm's
incentive to internalize imperfect markets when the cost associated with transacting
internally is lower. However, Coase also identified the tendency for firms to increase in
size and diversify geographically as the costs of decentralization (i.e., communication and
transportation costs) decrease, albeit in a purely domestic sense. Nevertheless, his work
ultimately led to the realization that the firm's boundaries are set in two dimensions:
product space (through vertical and horizontal integration), and geographic space (both
regionally and internationally).
Pointing once again to the dichotomy between the economist's definition of the
firm and the firm in "everyday speech," Phelan and Lewin (2000) argue that modern
theorists are really attempting to explain the existence of the "corporation." For them, a
corporation is broadly defined as an entity comprised of a number of people and other
assets, which may have legal status as a company or partnership. This rather broad
definition is a reflection of the structural diversity of the real-world organizations that are
referred to by modern economic theorists as firms.
The focus of more recent research "has shifted away from the coordination
problems originally emphasized by Coase and towards the role of firm boundaries in
providing incentives" (Holmstrom and Roberts 1998 p.74). Furthermore, strategists have
made significant contributions to understanding the benefits of the firm structure as well
as how resources affect their boundaries (Phelan and Lewin 1999). An extensive body of
literature is devoted to examining the relationship between firms and so-called hold-up
problems, transactions costs, and intellectual property rights. While much of the early
research was conducted in a domestic setting, Dunning (1977) provided a framework that
extends the analysis to the international allocation of economic activity.
Dunning's OLI Framework
The "Coasian" theory does not specifically address the emergence of the
multinational corporation. However, a more contemporary agenda of the strategic
management literature is to explain the determinants of foreign production, and hence,
the existence of multinationals. Within this area of research, Dunning's (1977)
Ownership-Location-Internalization (OLI) framework receives a disproportionate amount
In an effort to address why firms take ownership positions in foreign markets as
opposed to exporting or conducting transactions at arm's length, Dunning contended that
foreign direct investments (FDIs) are made when three conditions are fulfilled:
(1) the firm must posses ownership-specific advantages, those that are "internal to the
enterprise of the home country, but capable of being used with other resources in the
home country or elsewhere" (Dunning 1970 p.399). Ownership-specific advantages
include organizational and entrepreneurial skills, patents, and firm size (which may
lead to both scale economies and market power);
(2) the host country must posses location-specific advantages, those "originating only
from the resources of [the home] country but available to all firms" (Dunning 1970
p.399). Location-specific endowments include Ricardian-type endowments, i.e.,
proximity to the point of sale, market size, and availability of natural resources and
manpower; as well as the legal and commercial environment in which resources are
used, i.e., market structure, and governmental legislation and policies (Dunning
(3) and finally, there must be benefits to internalizing foreign production processes.
Addressing the two former conditions, Dunning (1980) states, "The possession of
ownership advantages determines which firms will supply a particular market, whereas
the pattern of location endowments explains whether the firm will supply that market by
exports (trade) or by local production (non-trade) [and hence FDI]" (p. 11). The third
condition addresses the mechanism by which firms exploit ownership and location
advantages in order to service foreign markets. In subsequent work, Itaki (1991) pointed
out that Dunning's early concept of internalization differs from that of Coase in that it "is
interpreted as internalization of an "ownership advantage" rather than that of an imperfect
market" (p.445). This is made clear in the following statement by Dunning:
The thesis is that the international competitiveness of a country's products is
attributable not only to the possession of superior resources of its enterprises but
also to the desire and ability of these enterprises to internalise the advantages
resulting from this possession; and that servicing a foreign market through foreign
production confers unique benefits of this kind. (Dunning 1977 p.402)
Nonetheless, Dunning subsequently reconciled his concept of internalization with that of
Coase, modifying it to include internalization of both imperfect markets and ownership
advantages (Itaki 1991).
Dunning (1980) later revisited his original argument and added that the desirability
of internalizing foreign production processes could derive from both market
imperfections and public intervention. Market imperfections include uncertainty in
future market conditions or government policies, structural imperfections (e.g., barriers to
entry, high transactions costs, etc.) and cognitive imperfections (e.g., unavailability or
costly acquisition of information about the product or service being provided), while two
types of public intervention were considered (at least, by Dunning) to be relevant to
The first concerns the extent to which government intervenes in the production and
marketing of public goods by corporations. Dunning (1977) states, "the need both to
generate innovations and ideas and to retain exclusive right to their use, has been one of
the main inducements for enterprises to internalize their activities" (p.404). In addition,
the efficient exploitation of technology often requires complementary resources that
cannot be protected by patent (e.g., financial systems, organizational skills, marketing
expertise and managerial experience). Dunning indicates that the lack of public
intervention in the production and marketing of these complementary resources also
encourages internalization by firms.
The second type of public intervention that promotes internalization is economic
policy which tends to distort the international allocation of resources. This includes
corporate taxes and policies regarding the remittance of dividends and other forms of
profit repatriation. For example, a multinational enterprise, in an effort to record profits
in the lowest tax areas, may find it desirable to control the prices at which intermediary
products are exchanged by its international groups. Thus, the firm may internalize
production across international boarders and use transfer pricing in order to avoid higher
sales taxes resulting from the exchange of intermediary goods at external-market prices.
However, Dunning recognized that as early as the 1970s, more intense governmental
surveillance over transfer pricing strategies had begun to erode some of these types of
benefits (Dunning 1977).
In his summarization of the OLI framework, Kurt Pederson (2002) provides a
useful figure that illustrates how the presence or absence of location and internalization
advantages might affect the mode of internationalization preferred by firms. In Figure 2-
1 (which is a reproduction of Pederson's figure), +I and +L indicate the existence of
internalization and location advantages respectively, while -I and -L indicate their
absence. In the figure, the firm's possession of ownership advantages is assumed.
Pederson isolates the role of location-specific advantages, suggesting that the
extent to which these types of advantages exist influences the firm's decision of whether
to make a direct investment or to serve the foreign market at arm's length. Loree and
Guisinger (1995) classify aspects of the policy environment (e.g., corporate tax rates,
(O advantages assumed) -L +L
-I (1) Simple export (3) Contractual agreements
+I (2) Sales subsidiary (4) Foreign direct investment
Reproduced from (Pederson 2002 p.6)
Figure 3-1. Modes of Internationalization
regulations on profit repatriation and foreign ownership, and investment incentives) as
location-specific disadvantagess and state that host-country governments have a
"normative desire...to manipulate policies that are thought to affect FDI flows" (p.285).
In his critique of the OLI framework, Itaki (1991) takes issue with the
inseparability of ownership and location advantages by pointing out that the eclectic
framework is "weakest when ascertaining which items are most decisive in attracting
FDI" (p.456). Using the case of a technological breakthrough made by a foreign affiliate
as an example, Itaki contends that the new technology would be classified by Dunning as
an ownership advantage of the parent MNC. However, to the extent that the cost of
developing and implementing the new technology is a product of local labor costs in the
foreign country, the advantage of the new technology (in economic terms) may be both
ownership-specific and location-specific. While Itaki's observations are valid, the
purpose of this study is not to make inferences on which of Dunning's three conditions
bears the greatest impact on the FDI decision. Rather, the usefulness of the OLI
framework lies in its taxonomical illustration of factors that may positively or negatively
affect the MNC's decision to invest in a given foreign country.
Foreign direct investment has been modeled by some researchers as the end-result
of a multi-stage process4. Although the context varies, the same analogy is useful in
setting a backdrop for this study. In the first stage, firms develop the initiative to
establish some form of international presence. This initiative could derive from
conditions specific to the home country, global economic conditions, or a mix of the two.
In the second stage, the firm's management makes decisions on the mode of organization
and location for its international operation. Here, it is argued that managers consider
location-specific attributes like the overall investment climate, market size and barriers to
trade (Dunning 1973). The third stage of the process involves making detailed economic
profitability forecasts given a prospective location for the foreign affiliate. Important
factors in this stage include the level of industrial concentration and existing competition
(Miller and Weigel 1972).
The literature review provided in this chapter is intended to provide insight into the
first stage of the process described above, while the empirical considerations of this study
pertain more directly to the second stage. Simply put, different aspects of the investment
climate are analyzed to determine their effect, if any, on the extent to which foreign firms
in general have found each country to be a suitable or unsuitable business location.
However, one should remain cognizant of the fact that a firm's decision to operate in a
foreign market does not necessarily result in foreign investment. Contractor and Kundu
(1998) provide clarification of this point by stating
Until a decade ago, local adaptation by global firms was expressed by varying their
business practices and methods in each country, while leaving the ownership and
organizational structures fairly invariant across nations. Today, the modal choice
issue has gone beyond the "internalize or not" question (Buckley and Casson,
4 See Miller and Weigel (1972) and Barrell and Pain (1996) for two examples.
1976), and even beyond the "licensing vs. joint venture vs. merger" set of
alternatives (in Buckley and Casson 1996), to include other types of [non-
investment based] alliances, such as management service contracts, and
franchising. The general modal choice set now includes varying levels of equity
ownership, as well as several alliances of various descriptions. The manager must,
today, choose from a larger set of options. (Contractor and Kundu 1998 p.353).
Thus, the literal intent of this study is to examine how firms respond to host-country
investment-climate conditions, with specific attention given to the affect on the decision
to participate via an equity-based mode of organization.
In an attempt to understand how the investment climate affects FDI decisions,
Stobaugh, Jr. (1969) specified the following basic approaches companies take when
analyzing the investment climate of potential host countries: (1) go-no go, (2) premium
for risk, (3) range of estimates, and (4) risk analysis. The two former approaches are the
least complex with regard to the required investigation and calculations, while the latter
two involve highly complex analyses. Nevertheless, it is useful to illustrate how
elements of the investment climate might affect the decisions made by firms who use the
two simpler approaches (i.e., go-no go and premium for risk).
For Stobaugh, managers who use the go-no go approach sometimes use it in
conjunction with the premium for risk approach and typically base the investment
decision on one or two characteristics of the host country. For instance, suppose the
manager is faced with a high probability that the host country currency will devalue, a
condition that directly affects the projected profitability of the venture. In this case, the
risk of devaluation might lead the manager to reject the investment alternative (in other
words, the manager chooses the "no go" option). On the other hand, if the probability of
devaluation is not high enough to warrant a rejection on the initial screening, the manager
might require a higher return on investment given the risk of devaluation, thus requiring a
premium for the risk. In either case, a host country facing a high probability of
devaluation is less likely to attract FDI than a country with a low probability.
This example provides an understanding of how just one element of the investment
climate (fluctuations in foreign currency values) might affect the outcome of the FDI
decision and can be extended to account for other investment climate factors as well as
firms who use alternative investment approaches. Other examples of location-specific
factors that could influence firms looking to invest in a specific foreign country might
include the level of political instabilityt, the nature of the corporate tax regime, the
extent to which foreign firms are protected from expropriation of assets and the level of
It is easy to imagine how each of these factors could lead a foreign firm to either
reject or accept an investment project on the basis of its perceived level of risk.
However, theoretical research into the foreign investment behavior of firms also suggests
that other factors often play an important role in the decision process. For instance, the
existence of resources with strong global demand in a particular country might attract
foreign investment irregardless of conditions that would otherwise translate into a
prohibitive level of risk for foreign firms. Alternatively, previous research has also
shown that the level of FDI may be related to the other types of capital (e.g., official
development assistance and private foreign lending) flowing into a given country.
The next chapter focuses on how empirical researchers have examined the factors
that affect the FDI decisions of multinational firms. Specific attention is given to the
strategic management stream of literature. Later, in Chapter 5, each of countries
examined in this analysis are analyzed for the extent to which other considerations might
dominate the effect of the investment climate.
The literature review presented in the previous chapter highlighted the distinction
between neoclassical economic theory and strategic management theory as each relates to
the direct investment activities of the multinational corporation (MNC). The purpose of
this chapter is to develop an empirical model that can be used to analyze FDI flows to
selected developing economies in the Western Hemisphere. The chapter begins by
reviewing several empirical studies specifically concerned with the determinants of
foreign direct investment as well as the factors that affect the mode of organization (i.e.,
ownership structure) and the location of FDI. The chapter ends with an in-depth look at
the variables that are analyzed in this study.
The MNC's Motivations for Making Direct Investments
It should be recognized that Dunning's OLI triad represents only one of many
attempts to understand why MNCs establish foreign operations. While some of the
alternative theories have gained impetus, most are derived from the same motivations and
often yield results that tend to reiterate Dunning's contentions. For example, in a direct
comparison of portfolio theory and the OLI framework, Morck and Yeung (1991) found
that the incentives for FDI derive from the benefits of internalizing foreign operations
rather than benefits associated with international diversification of corporate portfolios.
Morck and Yeung (1991) examined the extent to which multinationality affects a
firm's market capitalization (i.e., "market cap"), and thus the net worth of the firm's
shareholders. Assuming that financial markets operate efficiently, the MNC's market cap
(V) can be measured as the sum of its tangible (T) and intangible (1) asset values. Citing
a host of earlier studies (including work by Dunning, Vernon, and Modigliani and
Miller), the authors set out to determine the extent to which a firm's degree of
multinationality1 impacts its intangible asset value. Expenditures on R & D and
advertising were used to control for the effect of intangible assets like technical expertise
and consumer goodwill. The authors also control for different industries by including a
series of three-digit SIC2 code dummy variables.
Ordinary least squares (OLS) regression results indicated that the degree of
multinationality was positively correlated with the MNC's share price and that the
relationship was highly significant. On average, a firm with five or more subsidiaries
was found to have a share price 8.41% higher than that of a firm with less than five
foreign subsidiaries. More importantly, Morck and Yeung showed that the impact of
spending on intangibles like advertising or R & D also increased with the degree of
The implication of these findings, according to the authors, is that the value of
multinationality is not derived from the international diversification of risk, tax
advantages or relative production costs. Rather, the firm's possession of (ownership-
specific) intangible assets serves as a necessary condition for FDI to positively affect the
firm's market value. Thus, the results call into question the notion that investors value
1 The degree of multinationality is measured by the number of subsidiaries the firm has, the number of
foreign nations in which the firm has subsidiaries and a series of dummy variables representing various
levels of FDI (Morck and Yeung 1991).
2 Standard Industrial Classification
the MNC as a means for international portfolio diversification, and lend credibility to the
theory of internalization.
Pugel (1981) tested the effect of four ownership-specific variables the possession
of proprietary new technology, marketing ability and expertise, organizational and
managerial technique, and the ability to obtain capital at favorable rates on outward
U.S. FDI intensity, which was measured as the share of U.S. subsidiary profits in the total
after-tax industry profits in the host country. He used the proportion of scientists and
engineers in the total workforce as a proxy for the extent to which new technology is
generated through R & D. Advertising intensity in the industry served as a proxy for
marketing abilities and the importance of organizational and managerial technique was
measured by the share of total employment in the industry accounted for by managers.
Finally, the amount of capital necessary to establish a factory of minimum efficient scale
was measured by total assets (net of depreciation) multiplied by the location-specific
importance of scale economies (the measurement of which is explained below).
Using data roughly corresponding to the 3-digit SIC level on U.S. manufacturing
industries, Pugel found that all four ownership advantages favored outward FDI. He also
included two location-specific independent variables representing "centralizing" and
"decentralizing" agents. Namely, he examined the importance of scale economies in
production measured as the average size of the largest plants producing half of the
industry's output divided by total industry shipments and the magnitude of transport
costs. The four-firm concentration ratio (C4) of each industry was also included as a
measure of oligopolistic rivalry.
Pugel specified a double-logarithmic regression equation and used OLS to estimate
the relationships between the dependent and explanatory variables. As such, the resulting
parameter estimates can be interpreted as elasticities. The importance of scale economies
was shown to hinder outward FDI, as was the magnitude of transport costs, although the
effect of the latter was statistically insignificant. The C4 variable was positively related
to FDI, indicating that "oligopolistic reaction leads to more FDI than would otherwise
occur in an industry" (Pugel 1981, p.226). Pugel takes this finding as an indication of the
fact that firms operating in industries where production concentration is high tend be
more likely to protect their market share in foreign countries by establishing a subsidiary.
The findings of these studies suggest that a firm's FDI decision is impacted by the
need to protect proprietary information and expertise. However, the nature of this
relationship has been a source of debate among theorists. Along these lines, Lee and
Mansfield (1996) found that the adequacy of a developing country's system of
intellectual property protection (a location-specific variable observed by direct survey of
a random sample of U.S. firms) affects both the volume and composition of FDI in a
Least squares estimates on outward FDI data from 100 major U.S. firms in 14
developing countries suggest that "if the percentage of firms regarding protection in a
particular country as inadequate falls by 10 points, U.S. foreign direct investment there
might increase by about $140 million per year" (Lee and Mansfield 1996, p.185). Other
theorists have argued that a lack of intellectual property protection may favor FDI as
opposed to licensing or contracting. In contrast, the empirical results presented by Lee
and Mansfield suggest that a MNC's incentive to invest directly is positively related to
the level of intellectual property protection in the host country.
Decisions on the Mode of Organization
The study by Lee and Mansfield provides an example of how the scope of research
on FDI has broadened over time from analyzing the export versus FDI decision to
focusing on the incentives of internalization as opposed to licensing, subcontracting or
franchising. According to Buckley and Casson (1998), the 1990s witnessed a renewed
interest in why certain circumstances seem to favor some modes of market entry over
others. They make the following statement which, in fact, embodies the perspective from
which this study is approached:
Entry [into a foreign market] involves two interdependent decisions on location
and mode of control. Exporting is domestically located and administratively
controlled, foreign licensing is foreign located and contractually controlled, and
FDI is foreign located and administratively controlled. (Buckley and Casson 1998,
A review of the empirical literature on alternative modes of foreign entry and
ownership turns up studies on the decision between greenfield investment or mergers and
acquisitions (see Zejan 1990, and Hennart and Park 1993), joint ventures versus wholly-
owned subsidiaries (see various works by Contractor and Lorange, and Beamish and
Killing), and more broadly, the decision between foreign equity investment and non-
equity contractual alliances. Researchers have found that many of the same variables
play a role in each of these distinct decisions.
The theoretical model of entry strategies developed Buckley and Casson (1998)
yields the generalized result that high transaction costs cause firms to favor FDI over
subcontracting and licensing if the cost is associated with arm's-length technology
transfer, and over franchising if the cost is associated with the arm's-length intermediate
output market. However, empirical work by Contractor and Kundu (1998) indicates that
the choice of entry mode depends upon much more than transactions costs, pointing out
the relevance of agency costs as well as the importance of host country-specific variables
and firm characteristics. The authors provide an interesting study on the international
hotel industry, a sector in which foreign equity ownership is at least as widespread as
non-equity modes of organization (Contractor and Kundu 1998).
Contractor and Kundu specify their dependent variable as the mode of organization
and allow for four different modes: (i) fully owned, (ii) partially owned, (iii) management
or service contract and (iv) franchise. Host country-specific determinants included as
independent variables are Frost and Sullivan's "Composite Risk Index", the level of
foreign business penetration (measured by FDI/GDP), the level of development (i.e.,
GDP per capital and a measure of cultural distance between the home and host
countries3. The most interesting result of the analysis is that foreign equity-based modes
are less likely in risky countries and more likely in lower income countries.
The results reached by Contractor and Kundu were supported by survey
questionnaire responses reported by Kim and Hwang (1992), who surveyed 96
multinational managers on the importance of host-country environmental factors to the
decision between three modes of entry (i.e., licensing, joint venture and wholly-owned
subsidiary). The survey results indicated that perceived country risk and the level of
unfamiliarity with the host country are more important environmental factors than the
uncertainty of demand and intensity of competition in the host country. Specifically,
3 Firm-specific structural, strategy and control factors are also considered by Contractor and Kundu but are
less relevant to the intent of this study and are omitted for the sake of time.
higher levels of risk and unfamiliarity were associated with lower levels of resource
commitments and hence lower levels of FDI.
As part of the vast literature inspired by Dunning' seminal work, Agarwal and
Ramiswami (1992) examined the interrelationships between ownership, location and
internalization (OLI) factors and the choice of entry mode. The authors pointed out that
while there is a significant body of research devoted to examining the effect of OLI
factors, the interrelationships of the three types of variables have largely been neglected.
They concentrated on overseas leasing data from the U.S. equipment leasing industry and
specified the OLI triad as
* Ownership advantages such as firm size, multinational experience and the ability to
develop differentiated products
* Location advantages such as market potential and level of investment risk
* Internalization advantages such as contractual risk.
Overseas leasing is carried out by foreign financiers who either "export" the loaned
funds from the home country or establish a foreign leasing subsidiary. Furthermore,
foreign leasing subsidiaries are established through contractual arrangement, setting up a
new foreign entity, or by direct investment into a foreign leasing company. The authors
analyzed survey data consisting of responses from Presidents and CEOs of 97 U.S.
leasing firms on the effect of the OLI factors. The interacting relationships estimated by
Agarwal and Ramiswami were obtained from logistic regression and chi-square analysis.
The major findings were as follows: (1) large firms show a preference for
investment modes of entry in both low and high potential markets; (2) small firms with
limited multinational experience prefer the joint venture mode of entry in markets with
high potential; (3) firms with greater ability to develop differentiated products favor
foreign investment over exporting when contractual risks are high; and, (4) even in high
potential markets, substantial investment risk leads both small and large firms to export
instead of investing.
The results presented by Agarwal and Ramiswami highlight an important fact.
That is, although the empirical methods of each study mentioned above vary, they share a
common result. That is, host-country locational factors and their interactions with
ownership-specific factors play a key role in determining the MNC's choice of entry
mode (i.e., non-investment versus equity investment-based). The investment climate, or
at least investment risk, also appears to have a substantial effect on the decision of
whether or not to establish operations in a foreign location.
Decisions on Location
While many studies (including those cited thus far) have examined the effects of
ownership and internalization factors on the FDI decision, a large body of literature is
specifically devoted to analyzing the location decision. This is typically done by
modeling the cross-country distribution of FDI flows from a particular country (flows
from the U.S. and Japan have received a disproportionate amount of attention), groups of
countries (here, flows from the EU are the primary concern), or the entire universe of
countries. From a host-country's perspective, it is important to understand how firms
determine the geographic location of FDI. However, a randomly-selected group of
managers would likely point out a diverse collection of factors that affect this decision.
For instance, the manager of a labor-intensive firm may value low wage rates while the
management of a capital-intensive firm might tolerate higher wages in exchange for
lower interest rates. Nevertheless, analyzing the distribution of outward FDI flows does
provide some insight into how firms choose foreign investment locations.
Mody and Srinivasan (1997) analyzed the allocation of investment funds by
Japanese and U.S. investors by concentrating on the factors that caused investors to
differentiate between countries. They treat the FDI decision as a two-step process in
which the foreign investor first decides how much capital to invest abroad and then how
that pool of investment funds will be allocated across countries. Separate equations were
estimated for each supplier country, with the dependent variable specified as the share of
FDI host-country i received in each time period. The authors controlled for factors like
proximity in distance and methods of conducting business while testing for the effects of
location-specific variables like market size, the price of labor and capital, the corporate
tax rate, trade propensity, country risk, infrastructure, education and the existing stock of
FDI (the latter of which was included as a measure of persistence).
Several sets of regression coefficients were estimated. First, OLS estimators were
obtained, although the authors point out that these estimates are biased when unobserved
country effects are correlated with the observed explanatory variables. Second, the fixed-
effects model was used to eliminate the influence of unobserved country characteristics,
thus providing coefficients that "reflect [the] responsiveness of foreign investment to
changes within a country, over time" (Mody and Srinivasan 1997, p.784). Third,
between-estimators were provided with the intent to capture the variation in FDI shares
across countries4. While the authors recognize that each of these regression techniques
has the potential to result in biased estimators, they indicate that the results provide
different perspectives and can be used in conjunction to describe the variation in the data.
Finally, generalized least squares (GLS) estimators are obtained from a random-effects
4 Mody and Srinivasan (1997) interpret the fixed-effects estimators (or "within-estimators") as short-run
effects and the between-estimators as long-run effects.
model. The authors point out that the GLS estimates represent a weighted average of
within- and between-estimators, thus providing the best composite picture.
Country infrastructure and primary school enrollment rates (a measure of labor
quality) were found to be major influences on location only. Meanwhile, past investment
in a country affected both the timing and location of FDI. Corporate tax rates, the cost of
investment, country risk and wage inflation were not significant determinants of the
location decision, although the latter two did effect the timing of investment. Finally, the
authors found that the determinants of investment location for Japanese and U.S.
investors had converged over time.
As is the case with the model presented by Mody and Srinivasan, other studies on
the policy determinants of FDI location primarily concentrate on the effect of host-
country corporate tax rates. Loree and Guisinger (1995) indicate that this is primarily
due to the difficulty of capturing the incidence of other policy variables. However, data
compiled by the U.S. Department of Commerce and the Internal Revenue Service in 1977
and 1982 provided information on host-country investment incentives and performance
requirements, in addition to corporate tax rates. Loree and Guisinger (1995) tested the
effect of all three policy variables on the location of U.S. direct investment abroad. As
the authors point out, any empirical study that "attempts to model and observe the effects
of policies requires the inclusion of non-policy variables as controls of alternative
explanations" (p.285). Thus, they included GDP, infrastructure, and country risk as
independent control variables.
Loree and Guisinger made two interesting distinctions with regard to the
measurement of variables included in their model. First, the authors indicate that total
FDI flows include equity investment, reinvested earnings of foreign subsidiaries, and
other long- and short-term capital flows. Consequently, they specified the dependent
variable as the equity component of FDI, as opposed to total FDI flows. Second, they
point out that models intended to measure the response of FDI flows to changes in the
absolute level of the dependent variables often rely on the assumption that the stock of
FDI in each country is at equilibrium at the beginning of each period. Citing the
difficulty of maintaining such an assumption, the authors chose to measure the variables
in absolute levels.
Loree and Guisinger also include a dummy variable for the host country's status
as "developed" or "developing" (as classified by the OECD). They justify the inclusion
of this variable by pointing out that differences in the pattern of liberalization among
developed and developing countries suggest that the affects of policy variables in each
class may differ. So, in addition to testing the aforementioned relationships, the authors
tested for differential effects of policy variables among developed and developing
OLS coefficient estimates were reported for both policy and non-policy
determinants of U.S. FDI location in 1977 and 1982. While the results suggest that both
play a significant role, policy variables may be more important since they are able to be
altered quickly while non-policy variables take months, sometimes years to change
(Loree and Guisinger 1995). Although most of the non-policy variables had the expected
signs, only the effect of infrastructure was statistically significant in both years. The
authors indicate that the insignificance of the other non-policy variables is most likely
due to industry specific characteristics that can only be captured by segmenting the FDI
However, the interaction of the policy variables and the development-status
dummy provide some interesting inferences. First, host country performance
requirements had a negative influence on equity FDI flows to both developed and
developing countries. Second, investment incentives had a positive influence on equity
FDI in developing host countries only5. Finally, the most significant effect was that of
the corporate tax rate, which was negatively correlated with U.S. FDI in both periods
irregardless of development status.
Woodward and Rolfe (1993) concentrated on export-oriented FDI activity in the
Caribbean Basin over a four-year period in the mid-1980s. The data available to the
authors allowed them to examine the effect of location-specific characteristics on
manufacturing plant openings in the region. It was assumed that investors in the region
would choose one of 16 location alternatives by analyzing the level of profit their firm
could achieve by locating in that country. Using conditional-logit analysis, the authors
found that GNP per capital, the length of tax holidays and the likelihood of an exchange
rate devaluation each had a large, positive, and statistically significant effect on locational
probability. Significant negative relationships were found for wage rates, restrictions on
the repatriation of profits, inflation and transport costs.
5 Loree and Guisinger indicate that this may be due to competitive response from other countries ultimately
"ending in the prisoner's dilemma trap where all countries increase their [investment] incentives
simultaneously but no country increases its relative share of foreign investment" (1995, p.296).
The Empirical Model
The empirical analysis in this study draws upon existing research into the strategic
behavior of foreign direct investors while attempting to shed new light on the effect of
the investment climate on FDI flows to developing countries. The model is constructed
in such a manner as to include a set of location-specific factors that are considered by the
entities that make FDIs. Relevance to host-country policymakers in the Western
Hemisphere is considered paramount.
Some researchers have suggested that the locational characteristics that influence
foreign direct investors vary according to the intended purpose of the investment. For
market characteristics such as size, growth in size and income level are most
relevant for investment that seeks to access the host market. Other types of
investment, such as export-oriented, may be more concerned with other locational
characteristics such as wage rates or policies regarding export levels. (Loree and
Guisinger 1995 p.295)
In a study on FDI in developing host countries, Lecraw (1991) distinguishes
between three types of FDI natural resource-seeking, market-seeking, and export-
oriented efficiency-seeking. However, he suggests that despite the distinct characteristics
and motives for each type investment, there exists a set of common locational factors
(Lecraw 1991) that influence all inflows of FDI irregardless of the type6. Lecraw's view
is adopted here. That is, the range of factors considered in this analysis are intended to
represent a set of conditions faced by all foreign direct investors irregardless of whether
6 Lecraw (1991) indicates that these factors include: a change in the corporate tax rate, the growth rate of
the labor force, a change in the openness of the host country's policy toward FDI, a change in the country's
risk rating, a change in the real exchange rate, and the growth rate of infrastructure.
the purpose of investment is to gain access to natural resources, the domestic market, or
to establish an export platform.
Many researchers have analyzed bilateral investment data as opposed to investment
from the entire universe of source-countries. The most often cited reasons for using this
approach are that the researchers are examining the effect of distance between the home
and host countries (i.e., the "gravity" approach), or that it is hypothesized that the
relationships vary by the nationality of the investor. The gravity approach is not
appropriate for this analysis given that the study analyzes flows of FDI from the entire
universe of sources. Furthermore, empirical results presented by Mody and Srinivasan
(1997) suggest that the factors that drive the location of FDI have converged over time
for at least two of the most prominent sources of direct investment funds in the region:
the U.S. and Japan. Given these findings, it is reasonable to think that the use of
aggregate rather than bilateral investment data does not hinder the validity of the results.
Foreign direct investment flows are measured in dollars and "therefore depend on
the size of the economy" (Lecraw 1991, p.171). Meanwhile, the independent variables in
this study are (at least theoretically) independent of economic size. This makes it
necessary to standardize the dependent variable across countries. Addison and Heshmati
(2003) point out that the tradition in existing literature is to specify the dependent
variable as FDI as a percentage of GDP.
The Host-Country Investment Climate
Examining the effect that the host-country investment climate has on FDI
essentially requires the inclusion of three types of variables on the right-hand side (RHS)
of the regression equation. The first group consists of governmental policies and
regulations that affect foreign investment. Restrictions on foreign investment, foreign
ownership, and access to foreign exchange (i.e., barriers to profit repatriation) serve as
indicators of the host government's attitude toward foreign investment. Factors that
affect the return on capital to foreign investors make up the second group. Variables of
this type are the corporate tax rate and changes in the exchange rate. The third group is
comprised of measures of political stability and corruption.
Basi (1966) surveyed 160 U.S. firms via questionnaire and asked managers to rank
15 potential determinants of FDI location as either "crucially important," "fairly
important," or "not important." In Basi's survey, a majority of U.S. firms ranked
political stability in the host country and the government's general attitude to foreign
investment as "crucially important." It was also clear that other components of the
investment climate, although not weighted as heavily as these two, received a
considerable amount of attention from multinational managers. Table 4-1 summarizes
Basi's findings on the importance of several investment climate factors.
Table 4-1. Survey Results on the Importance of Investment Climate Factors
Number of Firms Ranking Determinant as:
Crucially Fairly Not
Foreign Country Investment Climate Factor Important Important Important
(1) Favorable Attitude Toward U.S. Investments 76 69 13
(2) Political Stability 101 58 1
(3) Tax Structure 30 101 27
(4) Stability of Currency Exchange 61 90 8
(5) Inflationary Trends 35 99 21
The Policy and Regulatory Environment
Governmental policy towards foreign investment has attracted a disproportionate
amount of attention over the last 10 to 15 years. In fact, Lecraw (1991) points to several
examples where changes in host-countries' attitudes toward foreign investment led to
dramatic changes in the magnitude of inward FDI (such as in China and Korea during the
1970s and 1980s). Several measures have been proposed in the literature to proxy for the
regulatory environment foreign investors face when conducting business in a given
country. In response to the need for transparent methods of measuring and comparing
regulatory regimes across countries, an increasing number of private and public
institutions have begun to provide numerical ratings of the regulatory risks foreign
The International Country Risk Guide (referred hereafter as ICRG) published by
the PRS Group encompasses one of the most comprehensive collections of data on the
economic, financial and political risks that investors encounter in developing countries.
The ICRG's composite political risk rating is comprised of twelve components measured
via subjective analysis of the available political information from individual countries.
Of the twelve components, the "investment profile" score provides the most relevant
measure of the regulatory environment facing foreign direct investors.
As defined by the ICRG (Sealy 2003), the investment profile is made up of three
risk factors: (1) the risk for contract viability/expropriation; (2) the risk for restrictions
on profit repatriation; and (3) the risk for payment delays. Each of the three factors is
assigned a score ranging from 0 to 4, with a score of 4 indicating "very low risk" and 0
indicating "very high risk." The three component scores are summed together to yield an
overall investment profile rating ranging from 0 to 12.
A score of 4 (the highest possible) for the first component is taken as an indication
that the host-country judicial system is likely to enforce contracts made between the
foreign investor and its domestic associates, and that there is little risk of unjustified
expropriation of assets by the host-country government. A high score for the second
component signifies that foreign parent-companies are relatively free to repatriate profits
and that the level of taxation on repatriated funds is relatively low. Finally, a high score
for the third component signals the existence of legislative protection against undue delay
of payment for services or products rendered. Using this methodology, the overall
investment profile score is expected to be positively correlated with FDI.
In order to better understand the ICRG investment profile rating, it is useful to
compare the cases of two countries in which the score differs significantly; Haiti had the
lowest investment profile score among the countries in the sample (at 5.5) in 2001, while
Uruguay had one of the highest (at 11.5). The inefficiency and lack of transparency in
Haiti's outdated legal system often hinder the resolution of disputes between foreign
investors and domestic parties. In fact, there have been more than 10 cases of
expropriation of private assets owned by U.S. interests in Haiti over the last 25 years.
While the Haitian government publically indicated its desire to alleviate these types of
situations, there has been little in the way of real action.
In contrast, property rights in Uruguay are recognized and protected by a well-
established and transparent legal system. In terms of dispute settlement, foreign investors
are usually given a choice of arbitration or court proceedings. Bankruptcy laws dictate
that creditors collect their debts first, followed by employees and then government.
There have been no instances of government expropriation in Uruguay in recent history.
Thus, it appears that the ICRG investment profile score does a good job of capturing the
actual state of affairs at least in the two example-countries examined here.
Political risk assessment has been "one of the fastest growing areas of research in
international business ... [as] the discipline has flourished in the wake of the
international turmoil of recent years" (Simon 1984, p.123). Assessing differences in
political risk levels across countries is a difficult process that often requires highly
specialized information and expertise. Thus, many MNCs have again turned to outside
sources for assistance in analyzing the political risks inherent to developing host-
countries. The academic literature is peppered with econometric models that employ
political risk indexes composed by a variety of public and private entities. Business
managers and academic researchers often choose the source that provides a measure that
is most relevant to their purposes.
This study examines two primary sources of political risk in developing countries:
government instability and corruption. With regard to the first, Butler and Joaquin
(1998) state, "Political risk is the risk that a sovereign host government will unexpectedly
change the 'rules of the game' under which businesses operate" (p.599). While it is
possible for an incumbent regime to have a "change of heart" with regard to its position
on foreign investment, it is more likely that political risk arises from an actual regime
change. Thus, government stability serves to measure the risk that the "rules of the
game" will change.
With regard to the second source, it is sometimes unclear whether corruption in
local government serves as "a beneficial 'grease,' a minor annoyance, or a major obstacle
for international investors" (Wei and Shleifer 2000, p.303). However, to the extent that
corruption results in so-called "crony capitalism," it is possible that corrupt government
practices at least encouraged the recent Latin American currency crisis by facilitating
"the misallocation of financial resources to the friends and relatives of government
officials" (Wei and Shleifer 2000, p.304) rather than to their most productive uses.
It is possible that even in the most moderate cases "financial corruption in the form
of demands for special payments and bribes connected with import and export licenses,
exchange controls, tax assessments, police protection, or loans may force the
withdrawal or withholding of an investment" (Sealy 2003, p.A-5). In the most extreme
case, the uncovering of corrupt governmental activities could trigger a "popular backlash
[requiring a major restructuring,] or, at worst, a breakdown of law and order" (Sealy
2003, p.A-6). For all of these reasons, corruption is viewed here as a detrimental
characteristic of the existing regulatory regime.
As previously mentioned, there are twelve components to the ICRG composite
political risk rating. Scores for government stability and corruption are two of the factors
included in the composite rating and are used here as measures for the perceived level of
political risk. The score for political corruption ranges from 0 to 6, with a higher score
corresponding to lower risk. This score is reached through a subjective assessment of the
extent to which the following conditions exist in the political system: "excessive
patronage, nepotism, job reservations, 'favor-for-favors', secret party funding, and
suspiciously close ties between politics and business" (Sealy 2003, p.A-6).
The ICRG's government stability score is a measure of the government's ability to
stay in office and carry out its declared programss. As such, the score has three
subcomponents: government unity, legislative strength and popular support. Each
subcomponent is given a value of 0 to 4, a higher score indicating a higher level of unity,
strength and support. The overall government stability rating is the sum of the three
scores. The ICRG's risk ratings for corruption and government stability are each
expected to be positively correlated with FDI.
With regard to the government stability rating, it is useful to compare the cases of
Argentina and Uruguay. The ICRG government stability score for Uruguay was 11.0 in
2001, while Argentina received a rating of 5.5. Uruguay had a stable democratic
government, and there have been no instances of political violence in the recent past. On
the other hand, Argentina was plagued by social tension and political unrest. Violent
protests led to the death of at least 25 Argentine citizens from 2000 to 2002. Once again,
the ICRG rating seems to provide a reasonably good representation of the conditions that
persisted in the two countries used as examples here.
Factors that affect the return on investment
The paper by Stobaugh, Jr. (1969) illustrates how managers use the projected
return on investment (ROI) from a foreign project to discriminate between investment
alternatives. As was discussed in Chapter 3, the projected ROI is sometimes used as a
basis for passing on an investment opportunity altogether (referred to as the go- no go
decision), while in other cases managers might require a premium for risk in countries
that are perceived as riskier investment locations (the example used in Chapter 3 was the
risk of currency devaluation). In either case, factors that negatively affect the projected
ROI decrease the host country's probability of attracting foreign investment.
Two variables that directly affect the return to foreign investors are considered in
this study. These ROI factors are intended to measure location-specific attributes of the
host country that affect income in two ways. First, the domestic (i.e., within the host
country) ROI is affected by the host-country corporate tax rate. Second, the foreign
owner's ROI is subject to changes in the foreign exchange rate. The following
paragraphs elaborate on these two points.
The effect of income taxation is straightforward; ceterisparibus, a one percent
increase in the corporate tax rate leads to a one percent decrease in the after-tax profit of
a corporation. While the corporate tax rate may not have a significant impact on the
competitiveness of an individual firm (given that all firms residing in the host country
are faced with the same taxes), it may be used by governments as a tool to attract FDI
that might otherwise be directed at another country. In fact, some researchers have found
evidence that OECD countries do compete with each other over corporate tax rates in
order to attract foreign investment. Furthermore, Loree and Guisinger (1995) indicate
that MNCs may choose to allocate a disproportionate amount of investment capital into
countries with the lowest tax rates.
The most commonly used measure for taxation in the empirical literature is the top
marginal corporate tax rate, but it is often the case that foreign firms face different levels
of taxation. However, the definition of a direct investment enterprise provided by the
World Bank (Liberatori 2003) requires that the firm be classified as a resident entity of
the host country. As such, it is assumed here that direct investment enterprises are taxed
in a manner comparable to a domestically-owned firm. The top marginal corporate tax
rate is used in this analysis as the measure of taxation and is expected to be negatively
correlated with FDI.
The effect of exchange rate regimes on FDI has attracted a significant amount of
attention in the empirical literature. A stable and predictable exchange rate has the effect
of allowing managers of a parent-company to be confident that they will be able to
transfer funds to and from its foreign affiliate without substantial risk of loss due to
exchange. Furthermore, "uncertainty caused by exchange rate volatility ... makes the
domestic value of foreign revenues and costs uncertain" (Darby et al. 2000, p.1). That is
not to say that domestic firms are not-at-all affected by the exchange rate. Rather, this
study is more interested in how it affects the attractiveness of a host country from the
vantage point of a foreign investor.
Goldberg (1993) points out three theoretical effects of exchange rate volatility on
foreign investment: (i) sectoral profitability effects; (ii) location effects; and (iii) wealth
effects. First, sectoral profits are subject to exchange rate-induced changes in product
demand and cost, thus impacting the international competitiveness of industries. Next,
exchange rate volatility alters the attractiveness of domestic and foreign production
locations, and hence domestic and foreign investment levels. Finally, the distribution of
wealth across countries is affected by movements in the exchange rate, which in turn also
alters the demand for domestic and foreign investments.
With specific attention given to the location effect, Aizenman (1992) found that
aggregate investment was higher in countries with fixed exchange rate regimes than those
with a flexible exchange rate. This finding led him to suggest that "the adoption of a
fixed exchange rate could encourage flows of foreign direct investment" (Aizenman
1992, p.913). It is important to note that Aizenmen made this finding under the
assumption that foreign investors are risk-neutral. Although, it is expected that
introducing risk-adversity would only compound the negative effect of exchange rate
volatility on foreign investment. Aizenman's finding, as well as those presented in
previous sections, lead to the expectation that exchange rate volatility will be negatively
correlated with net FDI inflows in the ensuing analysis.
In addition to testing the effects of the factors mentioned above, two additional
variables are included on the right-hand-side of the equation as controls for alternative
explanations. Although not specifically categorized as elements of the investment
climate, both income and the level of integration with foreign economies (also referred to
as "openness") are considered as traditional determinants of FDI flows (Addison and
Heshmati 2003). Following the precedent set in the existing literature, productivity is
measured by per capital GDP. Total trade (i.e., imports plus exports) as a percentage of
GDP serves as a proxy for a country's degree of openness or integration with the global
economy. Each of the control variables is expected to have a positive effect on FDI.
A linear relationship between the explanatory variables and the dependent variables
is assumed. Chapter 5 addresses the validity of this assumption and compares two
alternative models that may be used to estimate the data: the fixed-effects model and the
random-effects model. A series of test statistics are calculated and serve as the basis for
choosing the most appropriate method of estimation. Finally, the response of FDI to
changes in the investment climate is examined and discussed for each country in the
MODEL SPECIFICATION, ESTIMATION, AND EMPIRICAL ANALYSIS
Data and Sources
The data for this analysis were obtained from various sources. The foreign direct
investment share of GDP, total trade share of GDP and per-capita GDP series were
obtained from the World Bank's World Development Indicators 2003 CD-ROM database.
Corporate tax rate data came from the World Tax Database maintained by the University
of Michigan Business School1. Exchange rate data were obtained from the International
Monetary Fund's (IMF) International Financial Statistics website2. Finally, the
investment profile, corruption and government stability scores were obtained from the
PRS Group's International Country Risk Guide (ICRG) data wizard3.
Although data on FDI were readily available for many of the developing countries
in the Western Hemisphere starting in 1970, annual risk ratings were much more difficult
to obtain. In fact, the availability of these ratings limited the sample size to 21 countries
observed over the 18-year period 1984-2001. Gross domestic product data for Nicaragua
in 1999, 2000 and 2001 were estimated by increasing the 1998 value by the growth rates
reported by the Economist Intelligence Unit. The total number of observations in the
1 The World Tax Database compiles corporate tax rate information from a variety of sources and was
accessed at http://www.wtdb.org/.
2 The IMF's International Financial Statistics database can be accessed at http://ifs.apdi.net/imf/.
3 The ICRG data wizard can be accessed at http \ \ \ .countrydata.com/wizard/. Data are available for a
Table 5-1. Descriptive Statistics by Country (1984-2001)
FDI Investment Corporate Tax Rate Government
(% of GDP) Profile (%0) Stability
Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev
Argentina 1.962 1.911 5.750 1.517 30.500 5.102 6.806 2.023
Bolivia 3.987 4.264 6.472 2.841 25.833 1.917 6.611 2.725
Brazil 1.607 1.892 5.694 0.987 26.111 7.962 6.444 2.121
Chile 4.414 2.889 7.417 2.211 27.611 9.641 7.250 2.702
Colombia 2.010 1.121 5.750 1.517 33.111 3.341 6.556 1.917
Costa Rica 2.651 0.799 6.889 2.055 35.556 9.218 7.417 1.784
Dominican Republic 2.597 1.929 6.278 2.886 34.778 10.395 6.694 2.562
Ecuador 2.623 1.923 5.389 1.243 23.333 2.425 6.972 2.076
El Salvador 1.033 2.112 5.639 2.611 28.056 3.888 6.278 2.372
Guatemala 1.261 1.078 5.944 2.100 33.278 5.497 6.444 2.572
Guyana 8.018 9.945 6.056 1.697 48.889 6.978 6.250 2.451
Haiti 0.189 0.217 2.750 1.611 36.944 6.216 4.333 2.910
Honduras 1.787 1.208 6.194 1.582 32.222 9.111 6.306 2.408
Jamaica 2.836 2.412 6.639 2.127 35.253 4.485 6.917 2.302
Mexico 1.959 1.004 7.306 1.808 36.056 2.838 7.611 1.720
Nicaragua 3.430 4.383 4.472 1.398 33.694 5.396 6.944 2.363
Paraguay 1.183 1.006 7.139 1.747 30.000 0.000 6.667 1.749
Peru 1.899 2.380 5.861 2.071 34.444 8.024 6.417 2.451
Trinidad & Tobago 5.679 4.796 6.972 1.803 40.167 4.396 6.778 1.833
Uruguay 0.586 0.508 7.417 1.972 30.000 0.000 7.167 2.036
Venezuela 1.852 1.865 5.278 1.526 40.444 8.906 7.194 1.758
Sample 2.551 3.535 6.062 2.159 33.156 8.342 6.669 2.290
Exchange Rate Trade Per-capita GDP
Corruption Variability (%) (% of GDP) (US$)
Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev
Argentina 3.361 0.724 437.017 857.215 18.145 3.546 7,182.403 779.110
Bolivia 2.222 0.732 876.655 2,367.956 46.890 3.150 878.980 58.705
Brazil 3.500 0.618 463.218 558.691 18.466 3.492 4,296.506 235.791
Chile 3.333 0.485 14.669 13.037 59.003 4.790 3,959.624 1,059.139
Colombia 2.611 0.608 21.663 9.293 33.747 4.071 2,181.747 185.551
Costa Rica 4.889 0.471 12.495 4.386 78.088 10.789 3,234.244 419.880
Dominican Republic 3.222 0.548 23.659 33.641 66.303 5.991 1,575.407 249.774
Ecuador 2.972 0.436 46.817 24.846 56.924 7.165 13,519.929 4,008.081
El Salvador 2.750 0.809 8.900 16.008 53.900 9.336 1,524.806 180.581
Guatemala 2.611 0.916 13.913 17.555 40.560 6.677 1,423.258 89.060
Guyana 1.944 0.998 34.741 41.700 186.476 55.493 800.264 107.505
Haiti 1.333 0.840 9.781 14.476 39.201 8.287 433.151 69.767
Honduras 1.972 0.118 13.829 18.849 76.605 18.874 699.081 17.854
Jamaica 2.389 0.502 23.323 25.680 102.561 10.067 2,129.110 158.978
Mexico 2.889 0.471 36.308 39.946 44.711 14.202 3,315.535 234.241
Nicaragua 4.444 0.856 15,341.711 42,745.495 80.364 30.387 473.192 68.740
Paraguay 1.389 1.037 22.985 16.456 66.936 12.738 1,820.112 63.172
Peru 3.000 0.343 584.282 1,302.757 31.144 4.517 2,225.213 219.885
Trinidad & Tobago 2.833 0.383 6.097 8.208 83.601 14.255 4,468.905 436.402
Uruguay 3.000 0.000 45.462 29.240 41.069 3.169 5,639.432 768.806
Venezuela 2.944 0.236 37.774 25.338 48.480 7.111 3,472.989 142.421
Sample 2.839 1.041 860.729 9,661.129 60.627 38.776 3,107.328 3,057.545
resulting balanced data set is 3024. Table 5-1 presents some descriptive statistics on the
data used in the analysis.
Model Specification and Estimation
Panel estimates were obtained by regressing the dependent variable against a set
of explanatory variables for the 21 countries listed in Table 5-1. The basic regression
equation takes the form
Y = a + X,,, + E k Z,,, + E, (5-1)
where Y, represents "net inflows of FDI as a percentage of GDP" for country i (i=
1,2,...,N) in period t (t= 1,2,...,7); a is a constant term, and / and 0 are Kxl and
J x 1 vectors of unknown parameters to be estimated, respectively. The variable X
represents a vector of investment climate determinants of FDI and Z is a vector of
control variables, each of which varies in both country and time dimensions. The error
term ,, in the basic panel regression model has two components and can be expressed as
E,t = A, + V,t, (5-2)
where u, represents the variation unique to the cross-section and u,, is white noise.
Greene (2000) indicates that there are two frameworks commonly used to
generalize this basic model. First, the fixed-effects model, which is used to obtain
ii ithin-estimators, takes a as constant over time but specific to each country. The
dependent variable in the fixed-effects model is regressed against the difference between
the observed value of each of the explanatory variables and its mean, thus eliminating the
effect of unobserved country (or "fixed") characteristics. Thus, the fixed-effects
regression equation can be expressed as
[Y,, -,.]= a, + ,j ,[Xj,, x,.] + IOk[Zk, -Zk,.] + I, (5-3)
where ,. is the mean of country i 's FDI inflows, X,,. is a J x 1 vector of means on X,,
and Z,. is a K x 1 vector of means on Z,,, taken over the T observations respectively.
The intercept term (a, ) in this model varies by country and thus is an unknown
parameter to be estimated. Furthermore, a, is interpreted as the mean residual in each
country i and can be expressed as
a, = FDI.- J J XJ,. kq Z,. (5-4)
Equation 5-4 is unique to each country in that any change in the constant term represents
a parametric shift. The error term c,, in the fixed-effects model is defined as follows:
s, =Y,, -a, -Y,8,X X,, k Zk,,
=Y, -(,. -j,/jX, --k AZk.*) J8JXj,, -kAZkit
= [L, Y,]- I Y,,[X ,, X,.]- IY [Zk,, -Zk- ]. (5-5)
Second, the variance components model is used to obtain random-effects
estimators. These estimates are a weighted average of between- and ii ithin-estimators,
and thus reflect "both the influences across and within countries" (Mody and Srinivasan
1997, p.785). The intercept term in this model is not allowed to vary by country, hence
the estimated coefficients reflect the average country in the sample. Random-effects
estimators are derived from the following reformulation of the basic regression equation:
Y, = a + P,X,, + Zk Z,, +u, + (5-6)
In Equation 5-6, u, "is a random disturbance characterizing the ith observation and is
constant over time" (Greene 2000, p.568).
The random-effects model involves a transformation similar to the one made in the
fixed-effects model with one exception; when the mean is subtracted from each
observation, it is weighted by 1 A, where A is defined as
A C- (5-7)
(o2 + T-)
In Equation 5-7, ao is the variance of the basic error term (sE,), a2 is the variance of the
country-specific error term (u ) and Tis the number of years.
Some extreme cases should be explained purely for example. If the value of A
from the random-effects model is equal to 1, then the random-effects model is ordinary-
least-squares and the classical regression model defined by Equation 5-1 applies. At the
other extreme, a value of 0 for A would indicate that all of the variation in the data is
unique to the cross-section of countries. If this were the case, then the random-effects
model would be identical to the fixed-effects formulation.
Ordinary-least-squares (OLS) estimates were obtained first and serve as a basis
against which to compare the alternative methods of estimation. Following Greene
(2000), the first hypothesis test was for the existence of individual-country effects in the
data. The Lagrange multiplier (LM) test statistic4 from the least-squares regression took
a value of 66.3939, much larger than the critical value of 3.845. Thus, the null hypothesis
of no individual-country effects was rejected.
An F test comparing the fixed-effects estimators to those from the classical OLS
regression further suggest that individual-country effects exist in the data. The null
4 This statistic is part of the standard TSP output. A more detailed explanation of how the statistic is
computed can be found in Greene (2000).
5 Chi-square (95%)
hypothesis in this case is similar to the first LM heteroscedasticity test performed (i.e.,
there are no individual-country effects in the data, and hence there is no difference
between OLS and the fixed-effects model. The computed F statistic was 6.267, which
exceeds the 99% critical value of 1.88 for F(20,350). As such, there is strong statistical
support for the existence of country-specific effects.
Given the existence of individual-country effects in the data (and consequently, the
conclusion that the classical regression equation is inappropriate), the LM test statistic
based on the residuals from the fixed- and random-effects models was used to determine
the need to estimate heteroscedastic-consistent standard errors for each of the models.
The test statistics took values of 133.363 and 103.887 for the fixed- and random-effects
models, respectively, leading to a rejection of the null hypothesis of no heteroscedasticity
and indicating the need to compute heteroscedastic-consistent6 standard errors.
By estimating a set of robust standard error matrices for the fixed- and random-
effects models, it is possible to compare the two models and choose the most appropriate.
The preferred method is the Hausman test for correlation between the individual-country
effects and the other regressors in the equation (Greene 2000). If the individual effects
are correlated with the other regressors, the fixed-effects model is the most appropriate.
Conversely, if the hypothesis that the two are uncorrelated cannot be rejected, then the
random-effects model is the better choice. The Hausman test is based on the Wald
criterion and is asymptotically distributed chi-squared. The test statistic of 0.63254 was
considerably less than the 95% critical value of 2.167 for the chi-square distribution with
6 The ROBUST command in TSP performs this computation even if the nature of the heteroscedasticity is
the necessary degrees of freedom. Thus, the null hypothesis of no correlation could not
be rejected, and the random-effects model was deemed to be the most appropriate.
The Box-Cox transformation was made to several explanatory variables that were
hypothesized to have a non-linear relationship with the dependent variable. The
following transformation was made to the exchange rate, trade, corporate tax and per-
capita GDP variables:
x =- (5-8)
where the optimum value of 3 for each variable was estimated by scanning the range of
values between -2 and 2 (in increments of 0.1) and maximizing the log-likelihood
function of the fixed-effects regression equation. However, making the transformation to
these variables did not have a significant impact on the regression results and in each case
the linear model performed at least as well as the non-linear specification. Thus, the most
appropriate approach was to estimate the random-effects regression equation under the
assumption of linearity in the independent variables.
Results of Estimation
The parameter estimates and their associated standard errors can be seen in Table
5-2. Overall, the results suggest that the model performed well. The R2 value of .424
indicates that the right-hand-side (RHS) variables in the model explained nearly half of
the variation in the data. This can be considered a relatively high R2 given the size and
cross-sectional nature of the panel data set used in this analysis. Six of the seven
explanatory variables were statistically significant and only one took an unexpected sign.
Table 5-2. Random Effects Regression Results
Variable Coefficient Error t-statistic P-value
Investment Profile 0.00156060 0.00089697 1.679 0.093
Corporate Tax Rate (0.05498600) 0.02076900 (2.648) 0.008
Government Stability 0.00256093 0.00079045 3.240 0.001
Corruption (0.00019929) 0.00182247 (0.109) 0.913
Exchange Variability (0.00002477) 0.00001345 (1.842) 0.066
Trade 0.07202500 0.00599952 12.005 0.000
Per Capita GDP 0.00000160 0.00000084 1.890 0.059
Constant (0.03032100) 0.01162300 (2.609) 0.009
The investment profile score had a positive impact on FDI and was significant at
the 10% level. The parameter estimate of 0.0015 suggests that a one-unit improvement in
the investment profile score would have the effect of increasing the FDI share of GDP by
0.15%. Based on the definition provided by the ICRG (International Country Risk
Guide), the investment profile score takes into consideration the risks associated with
profit repatriation, contract viability and expropriation, and delay of payment. The score
would increase (or improve) if a the ruling regime in a country were to
* ease the restrictions on profit repatriation by allowing a greater percentage of
profits to be transferred out of the host country or by lessening the tax (or other)
penalties associated with such a transfer
* enact tougher legislation to ensure the enforcement of legal contracts and limit the
possibility of excessive delays in payment for goods or services rendered
* decrease the government's right to expropriate funds or other property belonging to
Thus, the positive relationship between the investment profile score and FDI is not
The government stability score was the most significant of the investment climate
variables examined. As defined by the ICRG, the government stability score is a measure
of government unity, legislative strength and popular support. The rating is intended to
provide an indication of the cohesiveness of the ruling regime and the extent to which
opposing parties pose a threat to its ability to remain in power. A one-unit increase in the
government stability score was estimated to have the impact of increasing the FDI share
of GDP by 0.26%. Thus, the results suggest that the potential for change in the host-
country regulatory regime is of substantial concern to foreign investors.
The amount of variability in the exchange rate had a negative effect on FDI,
although the magnitude of this effect was rather small. The exchange rate variable, as
specified in the model, does not reflect whether the change is due to an appreciation or
depreciation in the host-country currency. Rather, it simply measures the level of
stability in the local currency's value against the US dollar. Darby et al. (1999) showed
that exchange rate volatility leads to uncertainty in the future cash flows of a proposed
foreign investment project, and in turn, to a greater likelihood that the investment project
will either be postponed or rejected altogether. However, the results of this analysis
suggest that a 1% increase in exchange rate variability decreases the FDI share of GDP
by only 0.002%. It is noted that the risk of changes in the foreign exchange rate is a risk
that can be hedged in international financial markets (e.g., through the use of currency
swaps and futures contracts), and that this fact may account for the lack of response in
FDI to changes in the variable. The exchange rate variable was significant at the 10%
The parameter estimate for the corporate tax rate suggests that higher taxation is a
significant deterrent to foreign investment. Decreasing the corporate tax rate by 1% had
the effect of increasing the FDI share of GDP by as much as 5.5%. Furthermore, the tax
rate variable was significant at the 1% level.
The negative coefficient for corruption indicates that as the risk for political
corruption increases the FDI share of GDP increases. However, the variable was not
statistically significant. Nonetheless, the existing theoretical literature suggests that
corruption can serve as a benefit to foreign investors in the sense that they often posses
the capital necessary to "buy favors" from governmental agents. Although the regression
results seem to offer weak support for this contention, the most obvious inference is that
the level of political corruption does not appear to be a significant determinant of FDI in
the sample countries.
Finally, each of the control variables had a positive impact on the FDI share of
GDP. The trade variable was significant at the 1% level and the magnitude of its impact
was 0.072. This result suggests that countries that trade more also attract more FDI as a
share of GDP. Per-capita GDP was significant at the 10% level, although the magnitude
of its effect was much smaller. A $1,000 increase in per-capita GDP was estimated to
increase the FDI share of GDP by only 0.0016%.
Although the parameter estimates presented in Table 5-2 appear to be small in
magnitude, their corresponding effect on the absolute level of FDI is considerable in
many cases. This is important because it is the impact in terms of dollars with which this
study is ultimately concerned. As shown below, simple algebra can be used to reveal the
effect of the independent variables on the actual dollar level of FDI in each of the
countries in the sample.
The dependent variable is specified in the regression equation as the FDI share of
GDP for two reasons. First, as pointed out by Lecraw (1991), because the magnitude of
the explanatory variables is not dependent upon economic size, there is a need to
standardize the dependent variable (as defined in this study) across countries. This is
most often accomplished by dividing FDI by GDP. An alternative approach is to include
GDP as a discrete independent variable on the RHS of the regression equation. However,
the FDI and GDP variables were correlated with each other and GDP became the
dominant variable in the regression equation, detracting from the significance of the other
variables in the model. The regression equation was therefore estimated in shares and the
effect of each variable on the dollar level of FDI was then extracted from the model by
multiplying through by GDP as follows:
FDI,J/GDP, = a, + f/ X,,, + Yk Zkt (5-9)
FDI, =a, (GDP,t) + Xj,t (GDP,) + k k Zkt (GDP,t) (5-10)
Equation 5-10 shows that although the estimated parameters are constant across
countries, the magnitude of change in each country's level of FDI is unique to that
country. The remainder of this chapter is devoted to analyzing the relationship between
FDI and hypothetical changes in the investment climate variables in each country.
Equation 5-10 was used to create a benchmarking scenario where the parameter estimates
were applied to the most recent data available (2001). The implied impacts on FDI of
ceterisparibus changes to the explanatory variables in the model are presented for each
country in the analysis.
For the variables that consisted of a rating or score, the effect of a one-unit
improvement in the score is considered. For the corporate tax rate, trade and per-capita
GDP variables, the effect of a 10% increase is shown. Finally, the reaction of FDI to
eliminating any variability in the foreign exchange rate in other words, pegging the
local currency to the US dollar is considered. In each case, the change in the dollar