Embed Size (px)
Citation preview
Do determinants of FDI to developing countries differ
among OECD countries? Insights from a Bayesian
Panel Data Approach∗
Nikolaos Antonakakis† Gabriele Tondl‡
September, 2010
Abstract
The main objective of this paper is to examine the determining factors of bi-
lateral outward FDI stocks from OECD countries to developing countries located
in different world regions since the mid 1990s. Our goal is elucidate whether the
motivation for FDI differs among different OECD investors. Rather than relying on
specific theories of FDI determinants we examine them all simultaneously under a
unified framework. To achieve that, we employ a Bayesian Model Averaging (BMA)
approach in a panel data setting for 129 developing countries divided into 5 geo-
graphical regions for the 1995-2008 period. Among the main conclusions, we show
that OECD FDI outflows depends on market size and market growth, established
bilateral trade, openness of the host country, bilateral investment treaties, cultural
proximity, corporate taxes and the quality of institutions. The results also reveal a
rather heterogenous pattern of FDI determinants among the major OECD investor
∗Financial support from the OeNB Jubilaeumsfond, No. 11701, is gratefully acknowledged.†Vienna University of Economics and Business, Department of Economics, Research Institute
for International Economics and Development, Augasse 2-6, A-1090 Vienna, Austria, e-mail: niko-
[email protected].‡Vienna University of Economics and Business, Department of Economics, Research Institute for
International Economics and Development & Research Institute for European Affairs, Althanstrasse 39-
45, A-1090 Vienna, Austria, email: [email protected].
1
countries, which has important implications for the design of appropriate policies
in developing countries to attract FDI.
Key words: FDI determinants, Bayesian model averaging, Developing coun-
tries.
JEL codes: C11, F0, F21
2
1 Introduction
Since the mid 1990s OECD countries started to place an increasing share of their FDI
into developing countries (DC) in the regions of Europe & Mediterranean (EuMED),
Asia (ASIA), Sub-Saharan Africa (SSA) and Latin America & Caribbean (LAC). Their
presence in these countries reveals a distinct pattern.
If we regard the major four OECD investors (namely, US, German, French and Dutch
investors) we see that their presence in the above mentioned developing regions varied
substantially in terms of value and time. As an indicator, we consider FDI stocks held in
the host region per inhabitant of the investor country. For instance, according to Figure
1, in 1995, US FDI stocks per capita in Latin America and the Caribbean (LAC) and in
East and South Asia (ESA) amounted to 15.52 US$ and 8.88 US$, respectively, whereas,
in the developing regions of Middle East and North Africa (MENA), Europe and Central
Asia (ECA) and of Sub-Saharan Africa (SSA) accounted only for 1.65 US$, 0.88 US$
and 0.40 US$, respectively. In contrast, ESA and ECA captured the second and third
position, respectively, in favor of European FDI behind the developing region of LAC.
Despite growing concerns and political debate over rising protectionism in 2006, 40%
of the world share of aggregate FDI flows ($500 billion) in 2007 went into developing
countries (WIR, 2008). Even though the slowdown in the world economy and the financial
turmoil that led to a liquidity crisis in money and debt markets in many developed
countries and thus, to a decline in both inflows to and outflows from these countries, the
relatively resilient economic growth of developing economies has partly counteracted such
risk (WIPS, 2008).
In 2008, the equivalent amounts for US FDI stocks per capita in LAC, ESA, MENA,
ECA and SSA were 35.94 US$, 34.99 US$, 7.38 US$, 6.82 US$ and 1.30 US$, respectively,
as shown in Figure 1. Translating these number into the 14-year growth (between 1995-
2008) of US FDI per capita stocks gives a tremendous growth increase of 1.32, 2.94, 3.47,
6.73 and 2.25 for the respective regions as shown in Table 1, thus, leading to catching
up of US presence in ECA. In addition, the European countries increased their share
in all regions, with the particular increase in ECA, ESA and MENA. Furthermore, EU
investors increased their position in ECA and ESA more than the US and a few invested
3
over proportionately in SSA.
These facts raise several important questions: (i) what determines FDI flows from high
income countries to different developing regions? (ii) Given the ample of FDI determi-
nants considered in the literature, which are indeed the most crucial ones? (iii) Are they
homogenous among distinct regions? (iv) Do different OECD investors show different
motivations and care for different location factors? Answers to such questions are of great
importance for the initiation of FDI in specific regions.
A bulk of literature investigating the determinants of FDI has emerged aiming pri-
marily to pinpoint which factors FDI recipients have to provide to secure FDI inflows (for
instance, Wernick et al., 2009; Azemar and Desbordes, 2009; Bellak et al., 2008; Campos
and Kinoshita, 2008; Bevan and Estrin, 2004, for a review see Blonigen (2005)). In most
studies the key determinants of FDI like market potential for horizontal FDI and labor
costs for vertical FDI and distance are considered. Others emphasize the role of certain
types of FDI determinants like taxation, human capital, infrastructure, macroeconomic
factors, institutional factors and trade liberalization.
Several studies looked at the determining factors of specifics regions (for example,
Barrell and Pain (1997) for Eastern Europe, Trevino et al. (2008) for LAC, Asiedu (2006)
for Africa and Hattari and Rajan (2009) for ASIA). In contrast there are few studies that
look at the motives of the investor countries (see, for example for the US Nasser (2007)
and Udomkerdmongkol et al. (2009), for Germany Toubal et al. (2003) and Buettner and
Ruf (2007), for France Pfister and Deffains (2005) and Fontagne and Pajot (1998)).
Since the empirical FDI patterns reveal that the US is investing on a global scale while,
some European countries are clearly favoring certain regions, we think it is imperative
to learn how the motives of these investor differ. Moreover, whether these motives differ
between certain host regions. Therefore, this paper will systematically distinguish between
different investors and recipients. This is one of the major contributions of this paper.
Furthermore, although the literature has emphasized particular groups of determi-
nants, there is no study which looks at the complete set of determinants and indicates the
crucial ones. We will look at a large set of more than 30 potential determinants including,
market potential, labor costs, human capital, infrastructure, trade relations, macroeco-
4
nomic factors and institutional factors. Given the large number of potential determinants,
we apply Bayesian Model Averaging to assess which are indeed the robust factors for our
investor countries in different destinations. This is the second major contribution the
paper which has to offer.
The remainder of the paper is organized as follows: Section 2 discusses the determi-
nants of FDI and the hypothesis regarding our investor countries. Section 3 describes the
empirical methodology and the data. Section 4 presents the empirical results and section
5 concludes.
2 Determinants of FDI
In this section we present the theoretical and empirical evidence on the determinants of
FDI flows of the literature and propose the hypothesis concerning our investors.
2.1 Market size and Market Potential
Market size is one of the key determinants of FDI according to Dunnings Ownership,
Location Internalization (OLI) paradigm (Dunning, 1993a,b). Investing in large markets
gives access to large sales markets, and furthermore it permits to benefit from Economies
of Scale (EOS) (Amiti, 1998; Krugman, 1979). Thus, FDI flows are often explained in
gravity type models with the market size among the central variables. Empirically, the
relation between the host country’s market size and FDI is the most tested hypothesis
(Culem, 1988; Wheeler and Mody, 1992; Barrell and Pain, 1997, 1999; Bevan and Estrin,
2004; Bevan A. and K., 2004). Busse and Hefeker (2007) and Trevino et al. (2008) show
that market size is an important determinant of FDI in developing countries in standard
estimations. However, if the endogeneity of the regressor is correctly taken into account,
Busse and Hefeker (2007) and Campos and Kinoshita (2008) concluded that market size
and market growth are no longer significant determinants of FDI flows into developing
countries.
A positive relation between the host country’s market size and FDI suggests that FDI is
of the type of market-seeking investment (as opposed to efficiency-seeking and resource-
5
seeking investment which is discussed below). Not all countries invest in a region for
market-seeking motives, e.g. the US investment in Mexico is efficiency-seeking whereas,
Germanys FDI in Mexico is market-seeking (see, for instance, Vodusek, 2004). Investors
wishing to enter new markets prefer dynamic ones, i.e. markets with a good growth
performance. Furthermore, developed countries may prefer developed markets with a
higher GDP per capita over low income markets since they would meet higher demand
for the typical products of developed countries.
Consequently, we are interested to test whether market size, market potential and
market development is among the prime motive of our investor countries. Since all re-
garded investor countries offer leading world products we expect to find these to be major
motives.
2.2 Labor costs
Investment in developing countries often arises from the motivation to save labor costs
and thus, to dislocate a part of the production to low wage countries (efficiency-seeking
FDI). Therefore, we expect that FDI to developing countries is encouraged by high wage
gaps with the sender country. There are numerous examples for efficiency seeking FDI,
e.g. US investment in Central America, EU investment in Eastern Europe, developed
countries investment in the textile industry in East Asia. Konings and Murphy (2006)
found that that in the post-1992 period US-FDI in the EU-periphery was discouraged in
high labor cost countries. Bellak et al. (2008) estimated that increasing labor costs had a
negative effect on FDI inflows into the CEECs. Braconier et al. (2005) found that about
20 per cent of US multinational sales are based on low wages of skilled labor. We are
interested whether efficiency-seeking investment is an important motive for our investor
countries or if other types of investment rule. It might also be the case that investment
searches locations with relatively high wages since higher wages represent locations with a
higher productivity and better skills, required in branches like machinery and equipment.
In fact, the wage level is not the only labor related factor determining FDI investment
decisions. The availability of skilled labor and its productivity seem to be important for
firms as well. Filippaios and Papanastassiou (2008) looked at US multinational investment
6
in Europe and concluded that, besides labor costs, host country’s labor productivity in
the EU periphery is important for FDI inflows. Trevino et al. (2008) find that LAC
FDI inflows are related to educational attainment, measured by enrolment in tertiary
education. Azemar and Desbordes (2009) and Suliman and Mollick (2009) analyze FDI
flows to developing countries and conclude that the relatively low FDI flows into Sub-
Saharan Africa are partly explained by poor human capital and illiteracy. Noorbakhsh
et al. (2001) wonder why FDI flows to developing countries have reached only a limited
part of them. Their empirical analysis proposes that human capital is one of the most
important FDI determinants and has gained constantly in importance.
We look at both labor productivity and education in the labor force, proxied by
primary and secondary school enrollment to test whether these factors are important for
our group of investor countries.
2.3 Resources
Since the beginning of the 2000, resource-seeking FDI, geared by an increasing demand
and rising commodity prices for oil and minerals, boomed again (WIR, 2007). Thus,
the rising profits in the sector induced a flow of investment. Investment in extractive
industries involves large scale investment and high uncertainty of return. Investors can
act as monopolists. A good relationship with governments is essential. Autocratic regimes
and corruptive systems may facilitate operating the business. However, with this type of
the investment, political instability and the risk of expropriation can potentially lead to
high costs and losses. (WIR, 2007; Buckley, 2008).
Asiedu (2006) concludes that natural resources are besides market size the key deter-
minants for FDI in Africa.
We are interested in knowing to which extent our investor countries have placed their
FDI in resource abundant countries and which conditions destinations have to offer to
investors in extractive industries.
7
2.4 Host country’s trade openness, bilateral trade experience
and common trade policy framework
The embeddedness of the host country in international trade is relevant for FDI inflows
in several respects.
First, open economies -openness being indicated by exports plus imports over GDP-
are heavily linked with the world economy. They have liberal trade regimes, long estab-
lished international economic relations and are competitive on the world market. This
should provide a positive setting for investors. FDI would benefit from the liberal trade
regime which would facilitate to use of the affiliate as export base. Several studies find
a strong positive effect of openness on inward FDI. For different regions including East-
ern Europe, Asia, LAC and Africa, Campos and Kinoshita (2008); Trevino et al. (2008);
de Boyrie (2010); Sekkat and Veganzones-Varoudakis (2007) find that openness of the
host country is an important factor explaining FDI inflows.
Secondly, we propose that investors will have a stronger propensity to put their FDI
into countries with whom external relations have been already established. As evidence
we consider the position of the host country in the home country’s total trade over the
past 5 years.
Third, we conjecture that bilateral free trade agreements (FTA) encourage FDI, no-
tably efficiency seeking FDI. Bilateral trade agreements provide opportunities to dislocate
a part of, or the entire production in lower cost countries and to import the product with-
out trade barriers. There is evidence that FTA of the US with Central America has
generated important FDI flows into this region (Waldkirch, 2010). The same applies for
FTA between the EU and Eastern Europe (Baltagi et al., 2008). The perspective of ver-
tical FDI under FTA will increase with the wage gap of the host country, as argued in
Kim (2007).
2.5 Macroeconomic factors: exchange rate, inflation, debt
Macroeconomic stability has been stressed in numerous empirical investigations as an
important determinant of FDI (e.g. Campos and Kinoshita, 2008; Melanie Lansbury and
Smidkova, 1996; Asiedu, 2006). Macroeconomic stability involves low inflation rates, a
8
stable currency and low external debt. There are manifold examples that increasing ex-
ternal debts worsen the creditworthiness of countries, generate solvency problems and
lead to currency devaluations. Under these conditions, the investment can loose consid-
erable value. Currency devaluations and high volatility in exchange rates can also result
from current account deficits and other risk factors. High inflation and volatile inflation
increases uncertainty and thus, leads to higher investment risk. Consequently, FDI will
be discouraged by such conditions.
Busse and Hefeker (2007), Asiedu (2006), Campos and Kinoshita (2008) as well as
Trevino et al. (2008) stressed for our DC that the inflation level is an important factor
for FDI inflows. Serven (2002) proves that exchange rate uncertainty, i.e. volatility,
discourages private investment into DC. Clark and Kassimatis (2009) find that default
risk leads to FDI drops in LA.
We shall test the impact of exchange rate and inflation rate volatility as well as external
debt to find out to which extent these factors are relevant for FDI flows from our investor
countries.
2.6 Geographical and cultural proximity: distance, language,
colonial status
Middle European EU countries, such as Austria, have placed a large share of their FDI in
the CEECs. A large share of US FDI in EU-15 is located in the UK and in Ireland. Spain
has become a heavy investor in LAC. From these facts, we conjecture that geographical
distance, a common language and former colonial links promote FDI flows.
2.7 Institutional factors
In recent years the importance of institutions for FDI inflows has been increasingly
stressed. In view of the unequal FDI flows into Eastern Europe and LAC and the low
FDI record of Africa, institutions like the Worldbank underlined this factor.
The political system and quality of institutions is likely an important determinant of
FDI activity, particularly for less-developed countries for a variety of reasons. (i) Political
instability denoted by violence, civil war, or simply weak governments, will discourage
9
FDI. (ii) Countries with a developed democracy/political accountability provide a more
reliable legal base and therefore may encourage FDI inflows. On the other hand, Li and
Resnick (2003) argue that democratic countries limit Multinational Firms (MNF) to pur-
sue a monopolistic behaviour and local governments to offer generous incentives which
may reduce FDI. (iii) Poor legal protection of assets increases the chance of expropriation
of a firm’s assets making investment less likely. (iv) Poor quality of institutions neces-
sary for well-functioning markets (and/or corruption) increases the cost of doing business
(Antal-Mokos, 1998; Meyer, 2001) and thus, should also diminish FDI activity.
Poor institutions increase search, negotiation and enforcement costs thus, hindering
the establishment of new business relationships and the initiation of new transactions
(Antal-Mokos, 1998; Meyer, 2001).
In recent years the importance of these factors for FDI has been verified in a number
of empirical studies for large worldwide samples (e.g. Wernick et al., 2009; Busse and
Hefeker, 2007; Campos and Kinoshita, 2008), or specific regions Asiedu (2006) Naudand;
and Krugell (2007) for Africa, Barrell and Pain (1999) for Eastern Europe and Trevino
et al. (2008) for LAC.
Several international data sets are maintained which provide indicators of institutional
quality, among them the indicators of the International Country Risk Guide, of the Fraser
Institute and the World Bank’s World Governance Indicators. Since the latter provides
the largest coverage with respect to time and countries, we will use the indicators of this
data set.
2.8 Double Taxation Treaties and BITs
Since the early 1990s the number of Double Taxation Treaties (DTTs) and Bilateral
Investment Treaties (BITs) has grown significantly. BITs contain provisions for investor-
state dispute settlement with international institutions (e.g. the International center for
Investment dispute Settlement at the Worldbank) and reduce the uncertainty of expro-
priation (WIR, 2005). Our sample investors have concluded numerous DTTs.
Benassy-Quere et al. (2007) find that FDI inflows have become very sensitive to tax
rates in Eastern Europe and the enlarged EU respectively. Benassy-Quere et al. (2007)
10
and Bellak and Leibrecht (2009) argue that the tax elasticity of FDI is higher than with
respect to infrastructure. Big tax competition of countries wishing to attract FDI will lead
to an underprovision of infrastructure. Cleeve (2008) finds that fiscal incentives, namely
tax holidays are important for investors in Africa, Melanie Lansbury and Smidkova (1996),
Desbordes and Vicard (2009) investigated the impact of BITs and found that this
depends on the political relationship between the signatory countries. Only in case of
tense relationship, BITs promotes FDI flows.
2.9 Infrastructure
Foreign affiliates depend on the host countrys infrastructure in several aspects: They
wish to ship their manufactures or exploited products which requires a good transport
infrastructure. They have a need for communication with high technology media and thus
require a well functioning telecommunication and internet network.
Benassy-Quere et al. (2007) and Bellak and Leibrecht (2009) found that infrastructure
in Eastern Europe promotes FDI. More specifically, Campos and Kinoshita (2008) showed
that telecommunication is important for FDI in Asia and LAC and Bellak et al. (2010)
conclude that Information Computer Technologies (ICT) are an essential factor for FDI
in the enlarged EU.
3 Empirical Methodology and Data
The theoretical and empirical literature on the determinants of FDI has identified a large
number of variables as being correlated with FDI. A recent survey on FDI determinants
by Faeth (2009) presents nine theoretical models explaining FDI flows along with their
empirical performance. The author shows that there is no single theory of FDI, but a
variety of theoretical models attempting to explain FDI. In other words, not all deter-
minants in each of the nine theoretical models are found significant. Thus, any analysis
of FDI determinants should be explained more broadly by a combination of factors from
a variety of theoretical models. Put differently, the various FDI theories are typically
compatible with one another. For instance, a theoretical view holding that market size
11
matters for FDI is not logically inconsistent with another view that emphasizes the role
of openness on FDI.
Since theory does not provide sufficient guidance for selecting the proper empirical
model the issue of model uncertainty arises. So far the empirical literature has not at-
tempted to evaluate the robustness of FDI determinants. Model Averaging techniques
have been proposed to account for such model uncertainty. The basic idea behind model
averaging is to estimate the distribution of unknown parameters of interest across different
models. Two approaches can be distinguished: a) Model Averaging in the pure Bayesian
spirit (BMA) and b) Bayesian Averaging based on Classical Estimates (BACE). The fun-
damental principle of BMA is to treat models and related parameters as unobservable,
and to estimate their distributions based on the observable data. Based on prior infor-
mation on the parameters and considering all possible models, i.e. given by all possible
combinations of regressors, the posterior probability of models and regressors are esti-
mated. In contrast, BACE does not require prior information on all the parameters but
only the model size. It also considers all possible models and derives parameter estimates
by repeated OLS estimations and weighting models by their likelihood function.
Bayesian Averaging techniques have been applied in numerous empirical applications.
In the growth context, Sala-I-Martin et al. (2004) employ the BACE whereas, Fernandez
et al. (2001b) apply the BMA in a pure Bayesian spirit with different priors to deter-
mine the most robust growth regressors that should be included in linear cross-country
growth regressions. Leon-Gonzalez and Montolio (2004) extend the BMA to a panel data
framework.
We have a dataset on FDI determinants stretching over fourteen years and can thus,
investigate how the effect of FDI determinants changes across countries and over time.
Therefore we will apply a BMA in the panel data context. Since we regard bilateral
investment positions we do not encounter the issue of endogeneity.
3.1 Bayesian Model Averaging
Alternative models M j , with j = 1, ..., J , will be defined by the subsets of kj regressors
they include from the set of K regressors. Thus, all differ in their explanatory variables,
12
contain individual effects, alphai, and are linear regression models. Since it is assumed
that the individual effects enter in all models, the number of possible models is 2K . We
have data for N countries and T periods in each region.1 The dependent variables for all
countries and all models are grouped in vector y of length NT . The explanatory variables
and the N dummy variables for each country are stacked in design matrix X of dimension
NT×(K+N). β is defined as the full (K+N)-dimensional vector of regression coefficients
and individual effects. Any model M j with T observations for country i is represented
by:
yi = αiιT +Xji β
j + εi, (1)
where Xji is the Txkj submatrix of regressors of model M j and βj is the k vector of slope
coefficients, βj ε<kj (0 ≤ kj ≤ K). ιT is a column vector of T ones, and εi is the T×1 error
vector that is normal, with covariance matrix σ2IT , not autocorrelated and independent
of Xji , αi and βj. Although normality is not necessary for consistency, it guarantees
good finite sample properties (FLS 2001b). The effect of variables not contained in Xj is
assumed to be zero.
By averaging over all models the marginal posterior probability of including a certain
variable is simply the sum of the posterior probabilities of all models containing this vari-
able. Formally, the posterior distribution of any quantity of interest, say θj(= βj, σ, αi),
is an average of the posterior distributions of that quantity under each of the models with
weights given by the posterior model probabilities (PMPs):
p(θj|yi) =2K∑j=1
p(θj|yi,M j)p(M j|yi). (2)
This procedure is typically referred to as BMA and it follows from direct application of
Bayes’ theorem (Leamer 1978). P (θj|yi,Mj), the posterior distribution of θj under model
M j, is typically of standard form. However, we have to compute the PMPs due to model
uncertainty. Thus, we need to choose a prior distribution over the space M of all 2K
possible models. Following standard practice for BMA in linear regression models, we
allocate equal prior model probability to each model and set
p(M j) = 2−K . (3)
1For details, see TableA.2.
13
This yields a uniform distribution on the model space which implies that the prior prob-
ability of including a regressor is 12, which is independent of the combination of regressors
included in the model.2 With this prior model probability we get the following expression
for the PMPs:
p(M j|yi) =p(yi|M j)∑2K
i=1 p(yi|M i)(4)
where p(yi|M j) is the marginal likelihood of Model M j. This is given by
p(yi|M j) =∫p(yi|αi, β
j, σ,M j)p(αi;σ)p(βj|αi, σ.Mj)dαi dβ
j dσ, (5)
with p(yi|αi, βj, σ,M j) the model corresponding to 1, and p(αi, σ), and p(βj|αi, σ,M
j),
the parameter priors defined below in (6) and (7).
Computing the relevant posterior distributions is still subject to challenges as the
number of models to be estimated increases with the number of regressors at the rate 2K .
Furthermore, the derivation of the integrals implicit in (5) may be difficult because the
integrals may not exist in closed form. Using at least 30 regressors in our estimations, we
approximate the posterior distribution on the model space M by applying the Markov
Chain Monte Carlo Model Composition (MC3) methodology by Madigan and York (1995)
to simulate a sample from M. MC3 is based on a Random Walk Chain Metropolis-
Hastings algorithm which draws candidate models from regions of the model space in
the neighborhood of the current draw and then accepts them with a certain probability.
Posterior results based on the sequence of models generated from the MC3 algorithm can
be calculated by averaging over the draws.
The Bayesian framework needs to be completed with prior distributions for the param-
eters in each model M j: αi, βj, and σ. The choice of priors influences the results, which is
2There is some discussion about priors on the model space as many researchers prefer parsimonious
models. However, regular posterior odds ratios already include a reward for parsimony. Brock and
Durlauf (2001), among others, object to uniform model priors because of the implicit assumption that
a regressor’s probability is independent of the inclusion of others. They suggest a hierarchical structure
for the model prior. This, however, requires agreement on which regressors proxy the same theories. As
stated in Eicher et al. (2007b), such an agreement is often not existent and, thus, independent model
priors are preferable.
14
why non-informative priors would be preferable.3 However, PMPs cannot be meaningfully
calculated with improper non-informative priors for parameters that are not common to
all models. Thus, FLS (2001b) developed proper priors that do not require subjective
input or fine tuning for each individual model. Given their conclusions, we use the follow-
ing benchmark priors for our analysis. We take the {αi} to be independently uniformly
distributed on the real line and also adopt a uniform prior for the scale parameter common
to all models which gives us
p(αi, σ) ∝ σ−1. (6)
This prior implies that all values of α and σ for ln(σ) are given equal prior weight.
Furthermore, this distribution is invariant under scale transformations such as changes in
the measurement units. For βj we choose an informative g-prior structure
p(βj|αi, σ,Mj) ∼ N(0, σ2[gjX
′jXj]−1), (7)
with the following choice of the scalar hyperparameter gj
gj = min{ 1
NT,
1
(K +N)2}. (8)
This weighting factor depends only on the number of regressors and the sample size and
is a decreasing function of the latter. For the dataset considered in this study, this prior
resembles the one suggested by the risk inflation criterion (RIC) of Foster (1994) and has
a good small sample performance (Fernandez et al., 2001a).
3.2 Data
Our dataset consist of the major OECD investor countries namely, US, Germany, France
and Netherlands and of 129 recipient developing countries classified under four regions as
shown in Table A.2.
We are interested in the principal motives of OECD outward FDI stocks per capita
to developing countries and use a dataset of approximately 30 determinants of FDI for
the period of 1995-2008.4 Moreover, the construction of each of our variables is done in
3Two recent studies have analyzed the effects of prior choices in growth regressions regarding robust-
ness of parameter choices and coefficient estimates in detail (Ley and Steel 2009; Eicher et al. 2007b).4A detailed definition and sources of the variables is given in the Appendix in Table A.1.
15
such way to overcome possible endogeneity issues among our dependent variable, which
is outward FDI stocks per capita, and our explanatory variables.
4 Empirical results
We base our discussion below on the most important regressors having a Posterior Inclu-
sion Probability (PIP) above the recommended threshold of 0.50. According to Raftery
(1995), evidence for a regressor with a posterior inclusion probability from 50-75 % is called
weak, from 75-95 % positive, from 95-99 % strong, and > 99 % very strong. Masanjala
and Papageorgiou (2008) state that a PIP of 0.50 corresponds approximately to an abso-
lute t-ratio of one. Moreover, we discuss the regressors that are included in at least one
of the ten best models. These variables do not exert a robust effect themselves but are
relevant in combination with other regressors. Thus, they are relevant when it comes to
advocate policy packages instead of single policy measures.
We begin with the analysis of our estimation results for each of the four major OECD
investor countries to developing countries. Then we proceed by presenting the results for
the most robust FDI determinants to each of the developing regions as classified in Table
A.2. Our results of the BMA approach are based on the MC3 chain with 1.5 million
recorded draws (from which, the initial half million are discarded).
4.1 Determinants of FDI in developing countries
Table 2 reports the posterior inclusion probability (PIP), the posterior mean and the
posterior standard error of the BMA approach for FDI stocks to all developing regions for
each of the four OECD investors during 1995-2008. That is, we pool all regions together
in an attempt to unveil the broad determinants of OECD FDI in developing countries.
The results show that FDI to developing countries is above all market-seeking as
market size, GDP , is one of the most robust determinant for our OECD investors.
All European investors, but not the US, search FDI destinations with whom bilateral
trade relations, BTRADE, are established. Furthermore, practically all countries show
a strong preference for FDI destinations which are linked to them by a RTA. This might
16
be considered as evidence for vertical and efficiency-seeking FDI.
For European investors, FDI is importantly determined by the macroeconomic stability
in the destination, indicated by STDINF .
Institutional quality is not ranking generally as a robust FDI determinant. An insti-
tutional factor is only with two of them a highly robust location factor: Germany and
France highly care for government effectiveness, GOV ,5. On the other hand, poor quality
of Rule of Law, LAW ,6 and of Regulatory quality, REG, 7 are not an obstacle for most
OECD investors. US and Germany invest in developing economies that are growing slower
than the average, REL5GROWTH.
For German investors labor productivity, LPROD, and thus education of the work-
force, is a decisive factor. Productivity or wages are no robust determining factor with
the other investors.
Last but not least, German and Dutch FDI to developing countries is resource-seeking
(OIL and GAS, respectively).
Table 3 reports the posterior moments of the BMA approach with all regions pooled
and with the inclusion of corporate tax data. Since these data are not available for all
our countries throughout we restrict our country sample for which corporate tax data
are available. In addition to our previous results, which remain relatively similar, lower
corporate tax rates provide additional incentives only for EU investors in developing
countries.
4.2 Determinants of FDI in ECA by OECD investor
Having evaluated the broad determinants of FDI in developing countries we continue the
analysis of FDI determinants in each of the five developing regions.
5Government Effectiveness captures the perceptions of the quality of public services, the quality of the
civil service and the degree of its independence from political pressures, the quality of policy formulation
and implementation, and the credibility of the government’s commitment to such policies.6Rule of Law captures the perceptions of the extent to which agents have confidence in and abide by
the rules of society, and in particular the quality of contract enforcement, property rights, the police, and
the courts, as well as the likelihood of crime and violence.7Regulatory quality captures the perceptions of the ability of the government to formulate and imple-
ment sound policies and regulations that permit and promote private sector development.
17
Table 4 reports the posterior moments of the BMA technique for FDI flows to the
developing region of Europe and Central Asia (ECA) by each of our four OECD investor
countries. Despite the fact that OECD FDI to developing countries collectively is market-
seeking, the same does not apply for OECD FDI in the developing region of ECA. Only
US, German and French FDI in ECA prefer large markets, GDP , and are thus market-
seeking while Dutch FDI does not particularly concentrate on large economies. Clearly,
Dutch investors invest only in markets with a high labour productivity. With US and
French investors high labor productivity, competitive wages or a combination of both,
LPROD, and DWAGELPROD and DWAGE, play a certain role. In contrast these
determinants do not appear to be important for German investors. Thus, efficiency -and
indirectly education of the work force- is important to all investing countries except for
Germany. None of the investors exhibits a purely wage driven investment pattern.
Among all investing countries only France is to some extent resource-seeking, OIL.
For EU investors in ECA, established bilateral trade relations, BTRADE, and open
economies, OPEN , are important determinants. This confirms that EU investors focus
on countries which have already become fairly open and which show trade integration
with the EU. In contrast, US investors do not care for these factors. For them, investing
in ECA serves to enter entirely new markets.
European investors, in general do not discriminate in ECA between investment in
neighboring or more distant destination. Only German investors seem to be reluctant
to conduct FDI with distant countries, GDPDIST . The preference of Germany for
less distant investment is also mirrored by the coefficient of bilateral investment treaties,
BIT, which is -at first sight paradoxically- negative. Since Germanys FDI is located in
neighboring locations which are under the umbrella of the EU, it can renounce BITs which
ease investment in more risky destinations with no common legal framework.
Macroeconomic stability denoted by lower exchange rate fluctuations, STDEXC, is
an essential factor for US and German investors while,
Institutional factors show also a rather heterogenous pattern among EU and US in-
vestors in ECA. The US does not care or distinctly enters into institutionally risky mar-
kets, indicated by the negative sign of corruption, CORR, regulatory quality, REG, and
18
accountability, ACC.8 In contrast, the two big EU countries care for democratization,
ACC, and show a high like for destinations with high government effectiveness, GOV ,
while for the Netherland only government effectiveness, GOV , plays some role. German
investors care also for low corruption, CORR, and a sound legal situation, LAW . Thus,
above all Germany, and also to a considerable extent France, focus on institutionally de-
veloped countries in ECA. In other words, they invest in the advanced Eastern European
Countries which have entered EU. Those were requested to raise their institutional char-
acteristics by the EU. Thus, Germany and France are much less risk-taking in Eastern
Europe than US, and also to some extent Dutch investors. In addition to risk-takers, US
investors head to less developed ECA countries, DTEL (if we interpret the negative sign
in a way that there is poor advance in infrastructure development).
In summary, OECD countries have in general searched to invest in the large markets of
the transforming economies of ECA. FDI of the main European countries, Germany and
France, in ECA focuses on destinations which have provided relatively safe investment
areas due to their integration into the EU. The US, in contrast, shows a more distinct
pattern to explore new, and also more risky, investment destinations.
4.3 Determinants of FDI in ESA by OECD investor
Turning to the developing region of East and South Asia one can observe a rather distinct
pattern of FDI determinants.
According to the results reported in Table 5, all investors concentrate on countries
which have recently shown growth rates below the ESA average, REL5GROWTH. How-
ever, accounting for the fact that the least developed ESA countries show the highest
growth rates, this indicates that all OECD investors show a high preference for the more
developed ESA countries. Besides that, it appears that only Germany and France show
a preference for large markets which indicates that only their FDI in ESA is primarily
market-seeking.
8Voice and Accountability captures the perceptions of the extent to which a country’s citizens are able
to participate in selecting their government, as well as freedom of expression, freedom of association, and
a free media.
19
All investors search the more efficient destinations with better education, indicated by
the positive coefficient of LPROD. Simple wage competitiveness is no motivation.
European investors in general do not care for established trade relations, BTRA (the
Netherland is and exception here), and existing free trade agreements, RTA, which indi-
cates that their FDI strategy in ESA is to enter new, unfamiliar markets and serve local
demand. Nevertheless, a part or them (Germany, Netherlands) prefers locations which
have become open economies, OPEN . Thus, given the above reported characteristics, in
summary horizontal FDI dominates with EU investors. In contrast, US investors search
destinations with whom the US has a free trade arrangement, RTA. Together with the
insignificance of market size and the negative sign of BTRA, this suggests that US in-
vestors in ESA do not hesitate to enter unknown markets and a sizeable part of their
investment is purely vertical FDI, where goods are produced in more efficient locations
and shipped to the US under the RTA.
A considerable share of US, and some part of French and Dutch FDI in ESA is resource-
seeking as the abundance of gas, GAS, receives a positive posterior mean.
Macroeconomic stability is not a prominent factor for OECD FDI in this region, only
French FDI seems to avoid to some extent destinations with high exchange rate volatility,
STDEXC.
The impact of institutional factors in the host countries in ESA is rather contradictious
for OECD FDI. Thrivingly, all investors have a strong preference for destinations with
high government effectiveness, GOV . The US and Germany also search locations with
better regulatory quality, REG. For most investors, except for the Netherlands, a poor
quality of LAW is not discouraging. Furthermore, only US and German investors highly
care for democracy, ACC, while, French and Dutch do not. Only European investors
are discouraged by corruption, CORR, not the US. German investors are not avoiding
countries with little political stability, POL9.
In summary, investors prefer locations where a smooth functioning of business is guar-
anteed. Only US and German FDI is sensitive to a violation of democratic order.
9Political Stability and Absence of Violence captures the perceptions of the likelihood that the gov-
ernment will be destabilized or overthrown by unconstitutional or violent means, including politically-
motivated violence and terrorism.
20
Finally, the results show that US FDI is focusing on countries with slowly growing
telephone infrastructure, DTEL, which seems to confirm that the US locates in richer
ESA locations with an already abundant telecommunication infrastructure.
In summary, all OECD investors focus on the more advanced developing countries
in ESA which provide sufficient institutional effectiveness as business climate and show
a higher labour productivity. Macroeconomic factors are not decisive. US FDI in ESA
serves the dislocation of production into more cost-efficient location (vertical FDI) or to
exploit natural resources. German and French FDI is purely market seeking, focusing on
large, more developed markets.
4.4 Determinants of FDI in MENA by OECD investor
Table 6 reports the posterior moments of the BMA approach for FDI positions to the
developing region of MENA. The US, German and French FDI is operating in large
markets, GDP , while the Dutch is not. Thus, FDI of the big OECD countries in MENA
seems to be largely market-seeking.
For none of the investing countries established trade relations with the destination
seem to be decisive. France is even distinctly, and Germany to some extent, attracted
by locations with no trade relations, BTRADE. Thus, all countries invest in new, not
operated markets which probably have high import barriers. Nevertheless, the US strictly
prefers destinations with a strong presence of English when exploring such new markets.
Germany in turn prefers more globally oriented economies, OPEN .
US, French and Dutch FDI is not discouraged by poor labour productivity, LPROD,
which confirms the market-seeking nature of the investment. In contrast, German and
US investors are sensitive to the wage/productivity factor, DWAGELPROD, for the US
already a decreasing wage differential would be discouraging. In contrast, German FDI
would not be distracted by that. Given that there is no indication for a preference of
destinations covered under a bilateral RTA, except for French FDI, we conclude that FDI
in the regions is predominately horizontal FDI.
A share German FDI is also determined by the abundance in gas and is thus, resource-
seeking in nature. This type of investment is definitely not present with US investment
21
in the region, indicated by the negative sign of OIL.
Macroeconomic stability is no important factor for the OECD investors. Only Dutch
FDI avoids locations with high inflation volatility, STDINF , although it is not discour-
aged by mounting external debt, DDEBT .
With respect to the institutional framework, the US clearly shows a preference for
destinations with a democratic system, ACC. This is no factor for the other countries,
for Germany even explicitly no robust one. Besides democracy, US investors do not
care for institutional development and consider even a poor legal system, LAW , as no
impediment. The European investors do not seem to avoid FDI destinations in MENA
with poor institutional quality. Besides the non-consideration of Germany for democracy,
ACC, Dutch FDI shows a neglect for a sound legal situation, LAW , regulatory quality,
REG, and government effectiveness, GOV . Thus, in the institutional context, European
investment in MENA is more risk taking.
Finally, the negative sign of DTEL for US FDI would also suggest that the US invests
in more developed markets with an already developed infrastructure. Another minor
result is that a small part of German FDI is sensitive to the existence of DTT , while the
missing of a DTT is irrelevant for a some French FDI.
In summary, it appears that US FDI in the MENA region has searched for more
developed, more familiar and less risky countries in the region, to operate horizontal,
market-seeking FDI. European Investment is also primarily market-seeking, but also to
some extent resource-seeking (Germany) or vertical FDI (France). European investors
are much more willing to invest in politically and institutionally sensitive countries.
4.5 Determinants of FDI in SSA by OECD investor
According to Table 7 two striking facts appear: First, a large share of European FDI
in SSA is resource-seeking, OIL and GAS, which in contrast is not the case for US
investment. Second, Market size is only a significant determinant for French and Dutch
FDI, the latter entering only into large markets that were former colonies, GDPCOLON .
The US prefers destinations with high labor productivity, LPROD. Since the US is,
in addition focusing not on the fastest catching up economies, REL5GROW , in SSA, this
22
implies that it is focusing on the already more developed countries. In contrast, Germany
does not care for labor productivity in the destination and France and the Netherlands
invest even in regions with poor labor productivity. This implies that these investors
operate in poorer economies.
All investors prefer destinations with whom already an economic dialogue and re-
lations have established: For the US investing in existing trading partner countries is
important, Germany and France focus on countries which are tied to them by RTA, Ger-
many prefers destination which provide some risk-safety trough BITs, and the Netherlands
search former colonies. In addition, the US and France have negotiated DTT with their
FDI destinations.
Macroeconomic stability is only an issue for US FDI, which put a considerable weight
on low exchange rate volatility.
The preference of US FDI in SSA for less risky, more developed destinations appears
also in the high importance it gives to political stability, POL. This is the opposite with
European investors in SSA where institutional factors never reach a PIP above 50 per
cent. Germany cares to some extent for rule of law, LAW , and government effectiveness,
GOV , but some of its investment is not discouraged by corruption.
In summary, we find that European and US FDI follows distinctly different investment
patterns in SSA. The US is present in more developed, less risky destinations with whom
an established economic dialogue exists. US investment is not focusing on extractive
industries in SSA. On the opposite side, European investors heavily invest in resource
rich countries. Horizontal, market-seeking FDI is only found with French and Dutch FDI.
European investors are willing to enter poorer economies in SSA and do not explicitly care
for macroeconomic and institutional quality. Thus, it appears that European investors
are much more risk-taking than the US counterpart.
4.6 Determinants of FDI in LAC by OECD investor
Table 8 shows that all OECD FDI searches the larger LAC economies, GDP , and is thus,
to a large extent market-seeking. All OECD investors invest primarily in countries which
are not the most productive ones, LPROD. DWAGE or DWAGEPROD do not appear
23
as robust factors, suggesting that efficiency-seeking FDI and vertical FDI is subordinate.
For US FDI, the closeness of the market, GDPDIST , and the dynamic growth per-
formance of the destination, REL5GROWTH, is essential. In geographical terms, this
means that US FDI focuses on Central America, Columbia, Brazil and Argentina. France
equally has a strong preference for the higher growing countries of the region, while Ger-
many also seems to invest in the less dynamic economies.
Practically all investors search destinations with whom an agreement is maintained,
with the US, DTT , France, BIT , the Netherlands,RTA. Surprisingly, RTA does not
appear to be robust determinant for the US.
Macroeconomic stability does not appear as a highly important determinant, only
with Germany, SDEX is close to the 0.50 threshold. Neither are developed institutions
an important determinant for investment in LAC. German and French FDI do not mind
destinations with a poor government effectiveness, GOV . France is not discouraged by
corruption, COR.
In summary, our OECD investors mainly hold FDI in LAC by market-seeking mo-
tives. The US shows a strong preference for geographically closer LAC countries. Neither
macroeconomic stability nor developed institutions are important factors for our investors.
5 Conclusion and perspectives for further research
The purpose of this study was to shed light on the determinants of FDI in developing
countries. We looked at outward FDI stocks per capita from the four major OECD
investors into 129 developing countries and a set of 30 explanatory variables for the 1995-
2008 period. In an attempt to find robust explanatory factors and to account for specific
econometric issues, we estimated robust model specifications by Bayesian Model Averag-
ing (BMA). We look at around 30 different explanatory variables which can be clustered
into market size and market potential, labor costs, resources, openness, bilateral trade
relations and common trade policy framework, human capital, geographical and cultural
proximity, macroeconomic factors, double taxation and bilateral investment treaties, in-
stitutional factors and infrastructure. Finally, we allow for country heterogeneity across
countries in each developing region.
24
Generally, market size, established bilateral trade relations (with the EU investors)
and RTA (with US and France and Germany) were found as most robust determinants
of OECD FDI. Nevertheless, investor not only search for large markets alone, but in
some cases in combination with economies with less market potential. Macroeconomic
stability denoted by lower exchange rate and inflation volatility and the quality of in-
stitutions are main concerns for OECD investors. OECD investors care for both higher
labor productivity and labor costs and lower ones. Therefore, we do not find throughout
that OECD FDI is efficient-seeking. Common policy frameworks can also explain FDI
in several destinations. OECD FDI is also resource-seeking (especially, in SSA) as the
abundance of natural resources is an important determinant. Corporate tax reduction
provide additional incentives for EU investors in developing countries.
The results of this investigation can be regarded as work-in-progress.10 First, it is
surprising that no more factors generally stressed in the literature appear to be robust. To
elaborate on this point we shall do robustness checks such as using additional estimation
techniques and test threshold values by introducing interaction terms. Moreover, we shall
test whether a different sampling of destinations reveals new information.
References
Amiti, M., 1998. New trade theories and industrial location in the eu: A survey of evi-
dence. Oxford Review of Economic Policy 14 (2), 45–53.
Antal-Mokos, Z., 1998. Privatisation, Politics, and Economic Performance in Hungary.
Cambridge University Press.
Asiedu, E., 2006. Foreign direct investment in africa: The role of natural resources, mar-
ket size, government policy, institutions and political instability. The World Economy
29 (1), 63–77.
10Several factors constrain us at the moment to finalize the whole set of estimation. Above all, the
estimation of each specification requires about 8 weeks. Thus, estimations, amendments and further
results is time consuming. Apparently UK is among the major OECD investors. However, UK FDI data
are not publicly available and we will only be granted access at later time this year.
25
Azemar, C., Desbordes, R., Jun. 2009. Short-run strategies for attracting foreign direct
investment. Working Papers 24, Department of Economics, University of Glasgow.
Baltagi, B. H., Egger, P., Pfaffermayr, M., 2008. Estimating regional trade agreement
effects on fdi in an interdependent world. Journal of Econometrics 145 (1-2), 194–208.
Barrell, R., Pain, N., 1997. Foreign direct investment, technological change, and economic
growth within europe. The Economic Journal 107 (445), 1770–1786.
Barrell, R., Pain, N., 1999. Domestic institutions, agglomerations and foreign direct in-
vestment in europe. European Economic Review 43, 4–6.
Bellak, C., Leibrecht, M., 2009. Do low corporate income tax rates attract fdi? - evidence
from central- and east european countries. Applied Economics 41 (21), 2691–2703.
Bellak, C., Leibrecht, M., Riedl, A., 2008. Labour costs and fdi flows into central and
eastern european countries: A survey of the literature and empirical evidence. Structural
Change and Economic Dynamics 19 (1), 17–37.
Bellak, C., Leibrecht, M., Stehrer, R., 2010. The role of public policy in closing foreign
direct investment gaps: an empirical analysis. Empirica 37 (1), 19–46.
Benassy-Quere, A., Gobalraja, N., Trannoy, A., 04 2007. Tax and public input competi-
tion. Economic Policy 22, 385–430.
Bevan, A., Estrin, S., 2004. The determinants of foreign direct investment into european
transition economies. Journal of Comparative Economics 32 (4), 775787.
Bevan A., E. S., K., M., 2004. Foreign investment location and institutional development
in transition economies. International Business Review 13 (1), 43–64.
Blonigen, B. A., December 2005. A review of the empirical literature on fdi determinants.
Atlantic Economic Journal 33 (4), 383–403.
Braconier, H., Norback, P.-J., Urban, D., 2005. Multinational enterprises and wage costs:
vertical fdi revisited. Journal of International Economics 67 (2), 446–470.
26
Buckley, P. J., 2008. Do we need a special theory of foreign direct investment for extractive
industries? Journal of Chinese Economic and Foreign Trade Studies 1 (2), 93–104.
Buettner, T., Ruf, M., 2007. Tax incentives and the location of fdi: Evidence from a panel
of german multinationals. International Tax and Public Finance 14 (2), 151–164.
Busse, M., Hefeker, C., 2007. Political risk, institutions and foreign direct investment.
European Journal of Political Economy 23 (2), 397–415.
Campos, N. F., Kinoshita, Y., Jan. 2008. Foreign direct investment and structural re-
forms: Evidence from eastern europe and latin america. IMF Working Papers 08/26,
International Monetary Fund.
Clark, E., Kassimatis, K., 2009. The effect of country default risk on foreign direct in-
vestment. Economia Internazionale / International Economics 62 (3), 342–361.
Cleeve, E., 2008. How effective are fiscal incentives to attract fdi to sub-saharan africa?
Journal of Developing Areas 42 (1), 135–153.
Culem, C., 1988. The locational determinants of direct investments among industrialized
countries,. European Economic Review 32 (4), 885–904.
de Boyrie, M. E., 2010. Structural changes, causality, and foreign direct investments:
Evidence from the asian crises of 1997. Global Economy Journal 9 (4).
Desbordes, R., Vicard, V., 2009. Foreign direct investment and bilateral investment
treaties: An international political perspective. Journal of Comparative Economics
37 (3), 372–386.
Dunning, J., 1993a. The globalisation of business. Routledge, London, NY.
Dunning, J., 1993b. Multinational enterprises and the global economy. Addision-Wesley,
Wokingham.
Faeth, I., 2009. Determinants of foreign direct investment - a tale of nine theoretical
models. Journal of Economic Surveys 23 (1), 165–196.
27
Fernandez, C., Ley, E., Steel, M. F. J., 2001a. Benchmark priors for bayesian model
averaging. Journal of Econometrics 100 (2), 381–427.
Fernandez, C., Ley, E., Steel, M. F. J., 2001b. Model uncertainty in cross-country growth
regressions. Journal of Applied Econometrics 16 (5), 563–576.
Filippaios, F., Papanastassiou, M., December 2008. Us outward foreign direct investment
in the european union and the implementation of the single market: Empirical evidence
from a cohesive framework. Journal of Common Market Studies 46, 969–1000.
Fontagne, L., Pajot, M., 1998. Investissement direct a l’etranger et commerce interna-
tional: le cas francais. Revue Economique 49 (3), 593–606.
Foster, D., G. E., 1994. The risk inflation criterion for multiple regression. Annals of
Statistics 22, 1947–1975.
Hattari, R., Rajan, R. S., 2009. Understanding bilateral fdi flows in developing asia.
Asian-Pacific Economic Literature 23 (2), 73–93.
Kim, Y.-H., 2007. Impacts of regional economic integration on industrial relocation
through fdi in east asia. Journal of Policy Modeling 29 (1), 165–180.
Konings, J., Murphy, A. P., 2006. Do multinational enterprises relocate employment to
low-wage regions? evidence from european multinationals. Review of World Economics
(Weltwirtschaftliches Archiv) 142 (2), 267–286.
Krugman, P., 1979. Increasing returns, monopolistic competition, and international-trade.
Journal of International Economics 9 (4), 469–479.
Leon-Gonzalez, R., Montolio, D., September 2004. Growth, convergence and public in-
vestment. a bayesian model averaging approach. Applied Economics 36 (17), 1925–1936.
Li, Q., Resnick, A., 2003. Reversal of fortunes: Democratic institutions and foreign direct
investment inflows to developing countries. International Organization 57 (01), 175–211.
28
Masanjala, W. H., Papageorgiou, C., 2008. Rough and lonely road to prosperity: a reex-
amination of the sources of growth in africa using bayesian model averaging. Journal
of Applied Econometrics 23 (5), 671–682.
Melanie Lansbury, N. P., Smidkova, K., 1996. Foreign direct investment in central europe
since 1990: An econometric study. National Institute Economic Review 156, 104=114.
Meyer, K. E., 2001. Institutions, transaction costs, and entry mode choice in eastern
europe. Journal of International Business Studies 32 (2), 357–367.
Nasser, O. M. A., 2007. The determinants of the u.s. foreign direct investment: Does the
region matter? Global Economic Review 36 (1), 37–51.
Naudand;, W. A., Krugell, W. F., 2007. Investigating geography and institutions as de-
terminants of foreign direct investment in africa using panel data. Applied Economics
39 (10), 1223–1233.
Noorbakhsh, F., Paloni, A., Youssef, A., 2001. Human capital and fdi inflows to developing
countries: New empirical evidence. World Development 29 (9), 1593–1610.
Pfister, E., Deffains, B., November 2005. Patent protection, strategic fdi and location
choices: Empirical evidence from french subsidiaries&closecurlyquote; location choices
in emerging economies. International Journal of the Economics of Business 12 (3), 329–
346.
Raftery, A., 1995. Bayesian model selection in social research. Sociological Methodology
25, 111–196.
Sala-I-Martin, X., Doppelhofer, G., Miller, R. I., 2004. Determinants of long-term growth:
A bayesian averaging of classical estimates (bace) approach. American Economic Re-
view 94 (4), 813–835.
Sekkat, K., Veganzones-Varoudakis, M.-A., November 2007. Openness, investment cli-
mate, and fdi in developing countries. Review of Development Economics 11 (4), 607–
620.
29
Serven, L., Apr. 2002. Real exchange rate uncertainty and private investment in developing
countries. Policy Research Working Paper Series 2823, The World Bank.
Suliman, A., Mollick, A. V., 2009. Human capital development, war and foreign direct
investment in sub-saharan africa. Oxford Development Studies 37 (1), 47–61.
Toubal, F., Kleinert, J., Buch, C., 2003. Determinants of german fdi: New evidence from
micro-data. Discussion Paper Series 1: Economic Studies 2003,09, Deutsche Bundes-
bank, Research Centre.
Trevino, L. J., Thomas, D. E., Cullen, J., February 2008. The three pillars of institutional
theory and fdi in latin america: An institutionalization process. International Business
Review 17 (1), 118–133.
Udomkerdmongkol, M., Morrissey, O., Goerg, H., 2009. Exchange rates and outward
foreign direct investment: Us fdi in emerging economies. Review of Development Eco-
nomics 13 (4), 754–764.
Vodusek, Z., 2004. Foreign direct investment in latin america: the role of european in-
vestors. an update. SOE-IDB Working Paper Series 5, Inter-American Development
Bank.
Waldkirch, A., 2010. The effects of foreign direct investment in mexico since nafta. World
Economy 33 (5), 710–745.
Wernick, D. A., Haar, J., Singh, S., 2009. Do governing institutions affect foreign direct
investment inflows? International Journal of Economic and Business Research 1 (3),
317–332.
Wheeler, D., Mody, A., 1992. International investment location decisions. the case of us
firms. Journal of International Economics 33, 57–76.
WIPS, 2008. World Investment Prospects Survey, 2008 to 2010. United Nations, New
York and Geneva.
WIR, 2005. World Investment Report 2005: Transnational Corporations and the Inter-
nationalization of R&D. United Nations, New York and Geneva.
30
Table 1: Growth of FDI per capita between 1995-2008
Destination countries
Country of origin High-income OECD ECA ESA MENA SSA LAC
US 2.91 6.73 2.94 3.47 2.25 1.32
GER 3.46 16.42 8.15 38.43 1.58 1.79
FRA 5.88 19.98 7.40 19.09 6.75 2.59
NED 3.69 14.52 4.03 3.27 12.50 2.36
WIR, 2007. World Investment Report 2007: Transnational Corporations, Extractive In-
dustries and development. United Nations, New York and Geneva.
WIR, 2008. World Investment Report: Transnational Corporations, and the Infrastructure
Challenge. United Nations, New York and Geneva.
31
Figure 1: OECD countries FDI per inhabitant (in US$) position by region of destination
(1995) (2008)
0.88
8.88
1.65
0.40
15.52
3.974.33
0.77 0.74
8.32
1.65
3.29
1.88
1.15
5.77
7.01
22.54
6.33
1.42
17.58
0
5
10
15
20
25
ECA ESA MENA SSA LAC
US
GER
FRA
NED
6.82
34.99
7.38
1.30
35.94
69.19
39.66
30.20
1.92
23.23
34.60
27.64
37.68
8.88
20.72
108.77
113.41
27.04
19.15
59.08
0
20
40
60
80
100
120
ECA ESA MENA SSA LAC
US
GER
FRA
NED
Source: UNCTAD, National Bank Statistics and OECD.
32
Tab
le2:
Det
erm
inan
tsof
FD
Iin
dev
elop
ing
countr
ies
by
diff
eren
tin
vest
or
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
GDP
1.0000
0.1661
0.0119
BTRADE
1.0000
52.205
4.2383
GDP
1.0000
0.4145
0.0528
GDP
1.0000
0.1799
0.0701
REL5GROW
1.0000
-0.2871
0.0523
STDIN
F1.0000
-0.0976
0.0164
BTRADE
1.0000
63.746
9.0092
BTRADE
1.0000
40.837
3.1921
RTA
0.9994
0.0731
0.0147
RTA
1.0000
0.1494
0.0276
RTA
1.0000
0.2277
0.0324
STDIN
FL
0.8547
-0.0370
0.0188
LAW
0.2052
-0.0030
0.0064
GDP
0.9999
0.1806
0.0375
GOV
0.9983
0.0837
0.0169
GDPDIST
0.7410
-13.850
9.5045
STDEXC
0.0902
-0.0011
0.0039
GOV
0.9982
0.0719
0.0145
LAW
0.9403
-0.0541
0.0191
LPROD
0.2668
-0.0166
0.0298
GOV
0.0625
0.0011
0.0046
LAW
0.9236
-0.0434
0.0168
STDIN
F0.9350
-0.0706
0.0263
GAS
0.2311
0.0022
0.0043
DWAGE
0.0557
0.0078
0.0468
LPROD
0.8891
0.1043
0.0470
GDPLANG
0.2428
-0.0373
0.0712
GOV
0.1048
0.0020
0.0066
DWAGELPROD
0.0479
0.0005
0.0039
DTT
0.4454
0.0418
0.0509
REG
0.2305
-0.0102
0.0202
STDIN
FL
0.0335
-0.0005
0.0032
OPEN
0.3113
0.0061
0.0098
GDPCOLON
0.0988
-0.0127
0.0417
GDPCOLON
0.0326
0.0081
0.0503
REL5GROW
0.2543
-0.0852
0.1578
GDPDIST
0.0891
1.7901
6.2892
OIL
0.2438
0.0036
0.0069
DTT
0.0744
0.0068
0.0263
REG
0.1886
-0.0070
0.0158
BIT
0.0520
0.0028
0.0134
33
Tab
le3:
Det
erm
inan
tsof
FD
Iin
dev
elop
ing
countr
ies
by
diff
eren
tin
vest
or-
Res
tric
ted
sam
ple
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
GDP
1.0000
0.2233
0.0169
GDP
1.0000
0.3350
0.0766
GDP
1.0000
0.5117
0.0428
GDP
1.0000
0.2455
0.0678
ACC
0.8185
-0.0173
0.0097
BTRADE
1.0000
42.782
4.8002
BTRADE
1.0000
66.142
10.151
BTRADE
1.0000
37.816
3.6152
LAW
0.2853
-0.0060
0.0103
STDIN
F0.9995
-0.1245
0.0224
RTA
0.9938
0.1547
0.0365
GDPDIST
0.5098
-9.9394
10.724
RTA
0.1987
0.0087
0.0189
REL5GROW
0.9415
-0.5831
0.2123
STDIN
F0.9886
-0.1018
0.0250
GAS
0.3106
0.0037
0.0059
BIT
0.1092
-0.0042
0.0133
LPROD
0.9824
0.1831
0.0509
TAX
0.9828
-0.0825
0.0226
SDIN
0.1715
-0.0065
0.0154
DTEL
0.0996
-0.0056
0.0186
LAW
0.8397
-0.0518
0.0279
REG
0.6638
-0.0425
0.0346
STDEXC
0.0943
-0.0038
0.0131
REG
0.0694
-0.0012
0.0050
GOV
0.8188
0.0562
0.0320
GOV
0.2177
0.0119
0.0246
GDPLANG
0.0825
-0.0245
0.0908
REL5GROW
0.0617
-0.0091
0.0398
RTA
0.4977
0.0485
0.0536
LAW
0.2041
-0.0093
0.0201
TAX
0.0872
-0.0025
0.0089
LPROD
0.0484
-0.0021
0.0109
TAX
0.4670
-0.0262
0.0307
DIN
TR
0.0730
0.0013
0.0052
REG
0.4224
-0.0265
0.0342
LPROD
0.0671
0.0066
0.0275
DTT
0.3210
0.0364
0.0576
GDPDIST
0.2236
-4.0274
8.1691
34
Tab
le4:
Det
erm
inan
tsof
FD
Iin
EC
Aby
diff
eren
tin
vest
ors
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
GDP
0.9926
0.1504
0.0307
GDP
1.0000
1.3465
0.1103
BTRADE
1.0000
152.76
16.230
BTRADE
1.0000
124.75
6.7113
STDEXC
0.9921
-0.0482
0.0122
BTRADE
1.0000
38.670
5.7423
GOV
0.9481
0.1507
0.0515
LPROD
0.9923
0.1523
0.0302
DTEL
0.9809
-0.2535
0.0739
GDPDIST
1.0000
-303.07
27.728
GDP
0.9389
0.4668
0.1453
OPEN
0.3277
0.0910
0.1445
CORR
0.7672
-0.0392
0.0253
STDEXC
0.9975
-0.1675
0.0375
OPEN
0.9215
0.8235
0.3411
GOV
0.1499
0.0053
0.0141
DWAGELPROD
0.4442
0.0895
0.2391
ACC
0.9086
0.0759
0.0328
ACC
0.1853
0.0117
0.0277
REG
0.3640
-0.0191
0.0282
GOV
0.4124
0.0390
0.0522
OIL
0.1373
0.0050
0.0140
DWAGE
0.3403
0.0416
2.3170
OPEN
0.2384
0.1170
0.2321
LPROD
0.1337
0.0406
0.1188
ACC
0.2342
-0.0060
0.0121
BIT
0.1263
-0.0097
0.0290
DTT
0.0949
0.0091
0.0324
CORR
0.1206
0.0078
0.0240
DWAGELPROD
0.0903
0.0246
0.1553
REG
0.0780
-0.0053
0.0211
DWAGE
0.0826
0.1312
1.4723
LAW
0.0608
0.0029
0.0139
REG
0.0571
0.0038
0.0189
35
Tab
le5:
Det
erm
inan
tsof
FD
Iin
ESA
by
diff
eren
tin
vest
ors
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
REL5GROW
1.0000
-1.0952
0.2018
LPROD
1.0000
0.6848
0.1624
REL5GROW
0.9988
-2.0623
0.4603
LPROD
1.0000
0.4624
0.0658
LPROD
1.0000
0.5620
0.0368
REL5GROW
1.0000
-3.1894
0.3422
GOV
0.9873
0.1303
0.0370
REL5GROW
0.9930
-1.3185
0.3348
BTRADE
1.0000
-0.0718
0.0130
GOV
0.9981
0.0995
0.0223
LPROD
0.9055
0.6258
0.2789
GOV
0.9742
0.0766
0.0244
RTA
1.0000
0.3026
0.0422
ACC
0.9851
0.0581
0.0169
LAW
0.8783
-0.0706
0.0358
OPEN
0.9234
0.3693
0.1512
GAS
0.9959
0.0201
0.0048
LAW
0.9805
-0.0668
0.0202
GDP
0.4611
0.1526
0.1962
CORR
0.6335
0.0350
0.0307
DTEL
0.9854
-0.1779
0.0491
GDP
0.6405
0.1524
0.1304
GAS
0.3324
0.0083
0.0132
BTRADE
0.4513
4.4540
5.4992
GOV
0.9370
0.0465
0.0178
POL
0.4403
-0.0136
0.0172
CORR
0.2784
0.0213
0.0383
GAS
0.2751
0.0042
0.0076
LAW
0.5562
-0.0165
0.0167
OPEN
0.4243
0.1521
0.1972
STDEXC
0.1505
-0.0272
0.0733
ACC
0.5551
0.0140
0.0142
CORR
0.3003
0.0164
0.0279
DTEL
0.0886
-0.0155
0.0583
REG
0.1472
0.0040
0.0109
REG
0.2683
0.0128
0.0235
36
Tab
le6:
Det
erm
inan
tsof
FD
Iin
ME
NA
by
diff
eren
tin
vest
ors
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
GDP
0.9932
0.1988
0.0603
GDP
1.0000
0.7743
0.1189
GDP
1.0000
1.6751
0.0970
LPROD
0.9818
-0.1402
0.0405
GDPLANG
0.8490
0.2477
0.1313
REL5GROW
0.9999
-4.5610
0.9070
BTRADE
0.6239
-70.714
62.876
LAW
0.5095
-0.0392
0.0424
LAW
0.7177
-0.0321
0.0235
ACC
0.9989
-0.1662
0.0366
LPROD
0.3054
-0.0369
0.0620
REG
0.3514
-0.0246
0.0365
OIL
0.6832
-0.2067
0.1622
GAS
0.8509
0.3915
0.2094
DTT
0.2227
-0.0695
0.1456
GOV
0.1667
-0.0117
0.0294
LPROD
0.5968
-0.0620
0.0587
DWAGE
0.7651
-51.988
33.293
RTA
0.1577
0.0160
0.0420
STDIN
F0.1413
-0.0090
0.0254
ACC
0.5071
0.0166
0.0184
DWAGELPROD
0.7639
5.0553
3.2490
DDEBT
0.0992
0.0025
0.0088
DWAGE
0.4469
0.1847
0.3426
OPEN
0.4582
0.4409
0.5372
DWAGELPROD
0.3919
0.0112
0.0307
BTRADE
0.1437
-13.153
36.721
DTEL
0.3196
-0.0552
0.0902
GOV
0.1373
0.0131
0.0375
DTT
0.1182
0.0481
0.1517
37
Tab
le7:
Det
erm
inan
tsof
FD
Iin
SSA
by
diff
eren
tin
vest
ors
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
REL5GROW
1.0000
-0.5279
0.0561
RTA
1.0000
0.2829
0.0282
GAS
1.0000
0.0474
0.0071
GAS
1.0000
0.0311
0.0043
STDEXC
0.9996
-0.0285
0.0057
OIL
1.0000
0.0226
0.0023
GDP
0.9992
0.1931
0.0377
GDPCOLON
0.9993
0.5057
0.0958
DTT
0.9975
0.2126
0.0477
BIT
0.2808
0.0079
0.0139
RTA
0.9603
0.2926
0.0973
LPROD
0.1871
-0.0097
0.0224
LPROD
0.9823
0.0992
0.0209
LAW
0.2095
0.0031
0.0068
DTT
0.7909
0.3537
0.2172
POL
0.7951
0.0083
0.0051
CORR
0.1967
-0.0025
0.0055
LPROD
0.1236
-0.0133
0.0398
BTRADE
0.5248
0.0131
0.0139
GOV
0.1761
0.0030
0.0072
DTEL
0.0886
0.0074
0.0272
BIT
0.0943
-0.0050
0.0176
MIN
ORES
0.0730
-0.0103
0.0425
REG
0.0724
-0.0022
0.0091
OPEN
0.0487
0.0083
0.0439
POL
0.0458
0.0007
0.0039
38
Tab
le8:
Det
erm
inan
tsof
FD
Iin
LA
Cby
diff
eren
tin
vest
ors
US
GER
FRA
NED
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
PIP
Mea
nSE
GDP
0.9815
0.4409
0.2086
GDP
1.0000
0.5224
0.0514
OPEN
0.9994
-0.0290
0.0062
RTA
0.9997
0.2707
0.0599
LPROD
0.9807
-0.4020
0.1157
LPROD
1.0000
-0.4788
0.0835
LPROD
0.7375
-0.3974
0.2771
GDP
0.6980
0.1468
0.1247
GDPDIST
0.5346
-17.866
18.847
GOV
0.9950
-0.0847
0.0198
GDP
0.6954
0.2091
0.1586
OIL
0.6593
-0.0201
0.0166
REL5GROW
0.4597
0.2039
0.2469
STDEXC
0.4794
0.0321
0.0376
BIT
0.6768
0.0760
0.0613
LPROD
0.5045
-0.1401
0.1554
DTT
0.3054
0.0505
0.0855
REL5GROW
0.3214
-0.1139
0.1844
REL5GROW
0.6394
0.4190
0.3591
BTRADE
0.1412
-0.0086
0.0242
GOV
0.5100
-0.0396
0.0437
DDEBT
0.0974
-0.0004
0.0014
OIL
0.3565
-0.0115
0.0173
GDPLANG
0.0860
-0.0195
0.0744
CORR
0.3510
-0.0193
0.0292
POL
0.1160
0.0035
0.0113
39
A Appendix
40
Table A.1: Definition of variables
Variable Definition Source
FDI Annual log of outward FDI stocks per capita at current US$ UNCTAD, OECD and National Bank Statistics
GDP Log of GDP at constant 2005 international billions US$, PPP WDI
GDPPC GDP per capita at constant 2005 international US$, PPP WDI
REL5GROW 5-year average of annual growth rate of GDP at constant 2005 US$ Authors’ calculation based on WDI,
relative to 5-year average regional mean WEO and WIIW
DWAGE 1-diff of monthly wages of host country (at constant 2005 US$) Author’s calculations based on ILO,
as a share of home country’s monthly wages UNIDO, HDR and WDI
MINORES Minerals and ores exports (at current US$, SITC Rev.3, Authors’ calculations based on COMTRADE,
codes 27 and 28) as a share of total exports UNCTAD and WDI
OIL Log of Crude oil and NGL production (kt, kbbl/day) per capita Authors’ calculations based on World Oil
(data scaled upwards by 0.001 to avoid logs of zero values) Statistics IEA and Index Mundi
GAS Log of Natural gas indigenous production (thousand cubic metres) Authors’ calculations based on World Natural
per capita (data scaled upwards by 0.001 to avoid logs of zero values) Gas Statistics IEA and Index Mundi
OPEN Exports and imports (at current US$) divided by GDP at current Author’s calculations based on WDI
2005 international US$, PPP
BTRADE 5-years lag of exports and imports (at current US$) of home country to Author’s calculations based on IMF DOTS,
host country as a share of its total exports and imports of home country COMTRADE and WDI
RTA Dummy variable that equals 1 if there is Regional Trade Agreement into WTO
force between home and host country, 0 otherwise
BIT Dummy variable that equals 1 if there is Bilateral Investment Treaty into ICSID and UNCTAD
force between home and host country, 0 otherwise (date into force used)
DTT Dummy variable that equals 1 if there is Double Taxation Treaty between UNCTAD
home and host country, 0 otherwise
TAX Log of Highest marginal tax rate, corporate rate (%) WDI, KPMG, Michigan University, OECD, Price
Waterhouse, DoingBusiness, Eurostat, Tesche, WIIW,
IBFD, Deloitte, Central & East European Tax
Directory, Global Market Briefings, International
Tax Review, Investment Guide for Southeast Europe,
Ernst & Young,
LPROD Labor productivity defined as GDP per person employed at constant Authors’ calculation based on IMF-IFS, WEO, ILO,
2005 international US$, PPP WDI and United Nations Statistical Yearbook
DNETP 1-diff of log of Primary net school enrolment ratio (share) UNESCO, WDI, EDSTAT, UNESCAP and ECLAC
DNETS 1-diff of log of Secondary net school enrolment ratio (share) UNESCO, WDI, EDSTAT, UNESCAP and ECLAC
STDEXC Standard deviation of past 5 years of nominal exchange rate index Author’s calculations based on WDI, OECD IDIS
divided by mean of past 5 years of nominal exchange rate index and CIA World Factbook, various years
STDINF Standard deviation of past 5 years of CPI (2005=100) divided Authors’ calculation based on WDI, WEO, UN
by mean of past 5 years of CPI and CIA data
DDEBT 1-diff of external debt stocks (at current US$) as a share of total Authors’ calculations based on WDI, WRI, EIU,
exports (at current US$) UNECE, UNECA and CIA data
DIST Distance (in km) between sender country and destination country Author’s calculations based on CEPII
LANG Dummy variable that equals 1 if home and host country share a CEPII
common language (that is spoken by at least 20% of the population),
and 0 otherwise
COLON Dummy variable that equals 1 if home country was former colonizer CEPII
in the host country
ACC Voice & Accountability WGI
POL Political Stability & Absence of Violence/Terrorism WGI
GOV Government Effectiveness WGI
REG Regulatory Quality WGI
LAW Rule of Law WGI
CORR Control of Corruption WGI
DINT 1-diff of log of internet users (per 1000 people) WDI
DTEL 1-diff of log of telephone lines (per 1000 people) WDI
DWAGELPROD Interaction of DWAGE and LPROD
GDPDIST Interaction of GDP and DIST
GDPLANG Interaction of GDPE and LANG
GDPCOLON Interaction of GDP and COLON
41
Tab
leA
.2:
Cla
ssifi
cati
onof
Reg
ions
Reg
ion
nam
eC
ountr
ies
incl
ud
edN
um
ber
of
cou
ntr
ies
Eu
rop
e&
Cen
tral
Asi
aA
lban
ia,
Arm
enia
,A
zerb
aij
an
,B
elaru
s,B
osn
iaan
dH
erze
gov
ina,
Bu
lgari
a,
Cro
ati
a,
28
(EC
A)
Cze
chR
epub
lic,
Est
on
ia,
Geo
rgia
,H
un
gary
,K
aza
kh
stan
,K
yrg
yz
Rep
ub
lic,
Latv
ia,
Lit
hu
ania
,M
aced
onia
(FY
R),
Mold
ova,
Pola
nd
,R
om
an
ia,
Ru
ssia
nF
eder
ati
on
,S
erb
ia,
Slo
vak
Rep
ub
lic,
Slo
ven
ia,
Taji
kis
tan
,T
urk
ey,
Tu
rkm
enis
tan
,U
kra
ine
an
dU
zbek
ista
n
Eas
t&
Sou
thA
sia
Ban
glad
esh
,B
hu
tan
,B
run
eiD
aru
ssala
m,
Cam
bod
ia,
Ch
ina,
Hon
gK
on
g,
Ind
ia,
Ind
on
esia
,22
(ES
A)
Kor
eaR
ep.,
Lao
PD
R,
Maca
o(C
hin
a),
Mala
ysi
a,
Mon
goli
a,
Mya
nm
ar,
Nep
al,
Pakis
tan
,
Pap
ua
New
Gu
inea
,P
hil
ipp
ines
,S
ingap
ore
,S
riL
anka
,T
hail
an
d,
Vie
tnam
Mid
dle
Eas
t&
Nor
thA
fric
aA
lger
ia,
Bah
rain
,C
yp
rus,
Dji
bou
ti,
Egyp
t(A
rab
Rep
ub
lic)
,Is
rael
,Jord
an
,K
uw
ait
,18
(ME
NA
)L
eban
on,
Lib
ya,
Mal
ta,
Moro
cco,
Om
an
,S
au
di
Ara
bia
,S
yri
an
Ara
bR
epu
bli
c,T
un
isia
,
Un
ited
Ara
bE
mir
ates
,Y
emen
(Rep
ub
lic)
Su
b-S
ahar
anA
fric
aA
ngo
la,
Ben
in,
Bot
swan
a,
Bu
rkin
aF
aso
,B
uru
nd
i,C
am
eroon
,C
entr
al
Afr
ican
Rep
ub
lic,
41
(SS
A)
Ch
ad,
Con
go(D
emocr
ati
cR
epu
bli
c),
Congo
(Rep
ub
lic)
,C
ote
d’I
voir
e,E
qu
ato
rial
Gu
inea
,
Eri
trea
,E
thio
pia
,G
ab
on
,G
am
bia
,G
han
a,
Gu
inea
,G
uin
ea-B
issa
u,
Ken
ya,
Les
oth
o,
Lib
eria
,M
adag
asca
r,M
ala
wi,
Mali
,M
au
rita
nia
,M
oza
mb
iqu
e,N
am
ibia
,N
iger
,N
iger
ia,
Rw
and
a,S
eneg
al,
Sie
rra
Leo
ne,
Sou
thA
fric
a,
Su
dan
,S
wazi
lan
d,
Tan
zan
ia,
Togo,
Uga
nd
a,Z
amb
ia,
Zim
babw
e
Lat
inA
mer
ican
cou
ntr
ies
Arg
enti
na,
Bel
ize,
Boli
via
,B
razi
l,C
hil
e,C
olo
mb
ia,
Cost
aR
ica,
Ecu
ad
or,
El
Salv
ad
or,
20
(LA
C)
Gu
atem
ala
,G
uya
na,
Hon
du
ras,
Mex
ico,
Nic
ara
gu
a,
Pan
am
a,
Para
gu
ay,
Per
u,
Su
rin
am
e,
Uru
guay
,V
enez
uel
a(R
B)
Not
e:R
egio
ns
bas
edon
the
Wor
ldB
ank
clas
sifi
cati
on
42
