Economic and institutional determinants of FDI: an

application to the Portuguese case

Paulo Júlio∗ Ricardo Pinheiro-Alves† José Tavares‡

First draft: March 14, 2011

This draft: June 1, 2011

Abstract
This article analyzes the effects of several geographic, economic and institutional
factors on bilateral inward Foreign Direct Investment (FDI), from 45 developed
countries to 29 European countries. We conclude that geography affects to a
great extent bilateral FDI stocks. Market size and GDP per capita are the
most important economic determinants of FDI; however, the latter seems to
capture the effects of higher wages on FDI rather than the effects of a higher
purchasing power. Good institutions favoring economic freedom, low political
risk, and the ease of doing business are also shown to be key driving forces
of FDI. We also identify the institutions in which Portugal is lagging behind
the EU-27 and the EU-15 averages, and measure the benefits and the required
reform efforts of bringing them to the EU level. We find that the most promising
reforms consist in increasing the independence of the financial system, lowering
the levels of corruption, improving the strength and impartiality of the legal
system and the popular observance of the law, and improving some business
regulations—mainly by decreasing the bureaucracy to start a business and by
simplifying licensing procedures for some activities. Increasing labor market
flexibility to the EU level has also a large impact on inward FDI; however,
reforming this area comes at a very large cost.
JEL Classification: F30, H00
Keywords: FDI, Institutional reform, Institutions, Portugal, Europe, EU

∗
Gabinete de Estratégia e Estudos, Ministry of Economy, Innovation and Development, and
NOVA School of Business and Economics. Corresponding author: [email protected].
†
Gabinete de Estratégia e Estudos, Ministry of Economy, Innovation and Development, and
Instituto de Artes Visuais, Design e Marketing.
‡
Centre for Economic Policy Research and NOVA School of Business and Economics.
1 Introduction
The purpose of this article is twofold. Firstly, it analyzes the effects of economic and
institutional factors, including business regulations, on bilateral inward Foreign Direct
Investment (FDI), from 45 source countries to 29 European countries. To obtain a
full characterization of the institutional environment, we use 3 distinct databases: the
Index of Economic Freedom (IEF), the political risk rating from the International
Country Risk Guide (ICRG), and the Doing Business (DB) database. Secondly, it
evaluates the benefits of reforming Portuguese institutions to the European Union
(EU) average level—both for the EU-15 and for the EU-27. Our conclusions allows us
to formulate some policy recommendations as to the areas in which a reform should
be prioritized.
Since the 1990s, FDI has become increasingly important in a globalized economy,
both for developed and developing countries. According to the UNCTAD, FDI in
developing economies went up to around 28% of GDP in 2009, from 13% in 1990,
while in developed economies it has gone up to 31% from 9% of GDP in the same
time period. When registered in millions of dollars, this trend resulted in a worldwide
fivefold increase in FDI, at an yearly growth rate of 8.6% between 1990 and 2009.
However, developed economies are hosts of almost three quarters of these inflows, of
which more than half is targeted to the EU.
A large fraction of this trend can be explained by the increasing internationaliza-
tion of multinational firms. A well known model which seeks to understand why firms
may engage in FDI is due to Helpman et al. (2004). These authors consider that
firms rationally decide whether of serve foreign costumers, and whether they should
do so through exports or through outbound FDI. They conclude that only the effi-
cient firms choose to serve foreign costumers, and only the most efficient ones do so
through FDI.1 Whereas a firm deciding to export its products is able to concentrate
production in one place, benefiting from scale economies, it may have to incur in large
transaction costs, namely those associated with transporting and licensing a product
abroad, in order to successfully sell its products in the foreign market. FDI may
lessen these transaction costs, but also brings new ones, namely communication costs,
training costs for personnel, language barriers, and unfamiliarity with local business
and government practices. On the other hand, FDI may allow a multinational firm
to access cheap or qualified labor, natural resources and strategic assets, and better
1
Bhattacharya et al. (2010) extend this idea to the tradable services sector, by considering a setup
where transport costs are zero, but serving the foreign market through exports induces risk in the
consumer utility function. They conclude that, in this case, it is the least productive firms that
engage in FDI.

1
regulations and business practices (e.g. Dunning, 2008). Obviously, a corporation
should engage in FDI only when the net benefits of exporting a product to a given
market are outweighed by the net benefits of producing the product locally.
From the viewpoint of host countries, FDI also brings several advantages besides
its direct effects on investment and employment. Since FDI may be associated to
technology transfer, to the introduction of management skills, or even to improvements
in the productive structure of a country, it is more conductive to long-run growth and
to development than other forms of capital inflows (Borensztein et al., 1998; Barrell
and Pain, 1997). FDI may also have a positive impact on the balance of payments,
since multinational firms have a greater propensity to export than domestic firms.
It is therefore not surprising that a lot of research has been devoted to examine
and explore the determinants of FDI. A first wave of research articles focused solely
on its economic and geographic determinants, such as market size, growth, openness,
or the distance between countries. However, more recently, some articles concluded
that institutional and political risk factors also explain a large fraction of inward
FDI. There are several reasons why good institutions may promote FDI. First, good
governance is associated to higher economic growth, itself an important driving effect
of FDI. Second, good institutions prevent corruption and therefore decrease the cost
of investing abroad. Finally, good institutions foster political stability and decrease
political uncertainty (Walsh and Yu, 2010).
Hence, FDI is expected to flow to countries with a stable economic environment
and strong institutions, and where running a business is usually easier, ceteris paribus.
In the first part of this article, we show that good institutions favoring economic
freedom, low political risk, and the ease of doing business are key driving forces of
FDI. We also conclude that institutions favoring the absence of corruption and the
independence of the financial system tend to foster inward FDI. By the same token,
strong institutions which decrease a country’s political risk, mainly by contributing to
better socioeconomic conditions, to the strength and impartiality of the legal system,
to reduce the risks of expropriations, to eliminate the obstacles to the repatriation of
profits, and to increase the responsiveness of the government to its people, have also
a positive effect on inward FDI. The ease of starting, operating and closing a business
is also a relevant factor for FDI attractiveness.
This approach obviously leaves unanswered the question of which areas are most
promising for reform. In the second part of this article, we use three indicators to
measure the effects of reforming Portuguese institutions and business regulations to
the EU-15 and to the EU-27 levels, namely: the effect of the reform on FDI, the re-
quired reform effort, and the efficiency of the reform. We find that, for those areas in

2
which Portugal is lagging behind the EU average, the most promising reforms consist
in increasing the independence of the financial system, lowering the levels of corrup-
tion, improving the strength and impartiality of the legal system and the popular
observance of the law, and improving some business regulations—mainly decreasing
the bureaucracy to start a business and simplifying licensing procedures for some ac-
tivities. Increasing labor market flexibility to the EU level has also a large impact on
inward FDI; however, this comes at a very large cost.
This article is organized as follows. The next section reviews the literature. Section
3 presents the data used in the empirical analysis. Section 4 introduces the economet-
ric methodology. Section 5 discusses the results. Section 6 deals with institutional
reform in Portugal. Section 7 concludes.

2 Literature Review
The empirical literature on FDI is very extensive, but often very controversial. The
market size or market potential, usually a GDP measure, population, or economic
growth, are among the most extensively discussed economic determinants of FDI.
Billington (1999) and Scaperlanda and Balough (1983) find that market size and
growth have a statistically significant impact on FDI locations. Using a simultaneous
equation model, Tsai (1994) also shows that domestic market size is a key determinant
of FDI, but the role of growth is dubious. In another article, Janicki and Wunnava
(2004) show that GDP positively affects FDI for EU accession candidates. Other
studies which find a positive and significant relationship between market variables
and FDI include, among others, Love and Lage-Hidalgo (2000), Barrell and Pain
(1996), Wheeler and Mody (1992), Culem (1988) and Kravis and Lipsey (1982).
The literature also points out a positive relationship between FDI and host country
depreciations. For instance, Blonigen (1997) supports a positive relationship between
real dollar depreciations and Japanese acquisitions in the United States (US), as this
induces the acquisition of transferable assets within a firm across markets. A similar
result holds in Froot and Stein (1991), who present a story of capital market imper-
fections to argue that domestic currency depreciations can lead to foreign acquisition
of certain domestic assets. Kogut and Chang (1996) and Klein and Rosengren (1994)
obtain similar results.2 However, these studies rely solely on US data, which limits
the scope of analysis.
2
Swenson (1994) also addresses the relationship between exchange rates and FDI and reach similar
conclusions. The effects of exchange rate uncertainty in FDI are studied in Goldberg and Kolstad
(1995), Campa (1993) and Cushman (1985), among others.

3
 Taxes are also a key determinant of FDI. The analysis of the effects of taxation
on FDI dates back, at least, to Hartman (1984, 1985), who suggested a negative
relationship between taxes and FDI. Using a panel approach, Cassou (1997) also finds
that host country corporate tax rates have a significant negative impact on investment
flows. Related conclusions are also shared by Grubert and Mutti (1991), who show
that real investment responds to the host country effective tax rates, and by Devereux
and Griffith (1998), who show that average effective tax rates influence firm location
choices.3 However, contrary evidence is found in Swenson (1994), who finds that
greater average tax rates increase FDI. For an extensive review of this literature, see
de Mooij and Ederveen (2003).
Labor costs is another variable that is believed to have a negative impact on FDI
flows, ceteris paribus. This conclusion is supported, for instance, by Janicki and
Wunnava (2004), Bevan and Estrin (2004), Barrell and Pain (1996), Culem (1988),
and Flamm (1984), but others (e.g. Tsai, 1994, Wheeler and Mody, 1992, Kravis and
Lipsey, 1982) have found insignificant or opposite relationships. This mixed evidence
can be partially explained by the role of labor productivity in FDI, which is highly
correlated with labor costs.
The degree of openness has been studied, inter alia, by Bajo-Rubio and Sosvilla-
Rivero (1994) and Culem (1988). The former analyzes the determinants of FDI in
Spain, while the latter analyzes bilateral FDI inflows among 6 industrialized countries.
Both conclude that the degree of openness has positively affected inward FDI. Wheeler
and Mody (1992) and Schmitz and Bieri (1972) also analyze the impact of openness
on FDI, but find no statistically significant relationship. Culem (1988) and Schmitz
and Bieri (1972) also analyze the effects of trade barriers on FDI, with contradictory
evidence. While the former article reports a negative relationship, the latter posits a
positive one. Mixed evidence on the effects of trade barriers on FDI is also found in
Grubert and Mutti (1991).
Agglomeration effects are also relevant, as shown by Head et al. (1995), whose
conditional logit model suggests that agglomeration economies played an important
role in explaining Japanese manufacturing investments in the US. Devereux and Grif-
fith (1998) construct 3 measures of agglomeration and, using a nested multinomial
logit model, conclude that agglomeration effects influence the locational decisions of
US firms.
Some studies have devoted attention to the level of infrastructure. For instance,
3
Additional studies that yield similar implications include Billington (1999) and Hines and Rice
(1994). Hines (1996) finds that state taxes significantly influenced the pattern of foreign direct
investment in the US.

4
Loree and Guisinger (1995) use principal component analysis on 22 infrastructure
measures to reduce them to 6 variables, of which only 2 are retained in the empirical
specification. These reflect the amount of communication structure and the amount of
transportation infrastructure of a country. The authors then show that these compo-
nents have an important role in FDI. Bellak et al. (2007) follow a similar path and use
an augmented gravity model setting to show that telecommunication and transport
infrastructure play a role in the location decisions made by multinational enterprises.
Biswas (2002) uses main telephone lines per 100 inhabitants and per capita total net
installed capacity of electric generating plants, and finds that both have a statistically
significant impact on FDI. Wheeler and Mody (1992) conclude that agglomeration
economies influence investors decisions, and that good infrastructure development
makes investors less sensitive to short-run incentives.
Education has not been extensively addressed as a determinant of FDI. One ex-
ception is Walsh and Yu (2010), who find that education has a negligible or a slightly
counter-intuitive negative effect on FDI, depending on whether one considers FDI in
the secondary sector, or in services. Altomonte and Guagliano (2003), on the other
hand, find that education has a negative effect on a multinational’s probability to in-
vest in Central and Eastern European countries, or in Mediterranean countries if that
investment is made on traditional industries, but has a positive and significant impact
on that probability if the investment is made in the services sector in Mediterranean
countries.
Institutional factors are the main focus of this article. Schneider and Frey (1985)
were among the first to empirically address these explanatory variables by presenting
an inverse relation between political instability and other risk factors with incoming
FDI. More recently, institutional factors have fostered the research agenda on FDI.
For instance, Wei (2000a,b) concludes that corruption reduces inward FDI—firms or
individuals may be required to pay bribes to government officials in order to obtain
permits, licences, or other government services in order to run a business in a country,
therefore increasing the costs of doing business. Some recent studies (e.g. Lee and
Mansfield, 1996, Knack and Keefer, 1995) have shown that property rights and the
protection of intelectual property influence the amount and the composition of FDI.
Biswas (2002) shows that both traditional factors and nontraditional (institutional)
factors are important determinants of FDI inflows, and Stevens (2000) and Bénassy-
Quéré et al. (2007) show that political and institutional factors explain an important
part of FDI, which cannot be explained by economic factors alone. However, other
studies were not able to establish a relationship between FDI and institutional per-
formance. In particular, Bevan and Estrin (2004) find no significant impact of insti-

5
Table 1: Host and source countries.

Host countries Source Countries

Austria, Belgium, Bulgaria, Croatia, Czech Argentina, Australia, Austria, Belgium,
Republic, Denmark, Estonia, Finland, Brazil, Bulgaria, Canada, China (except
France, Germany, Greece, Hungary, Ire- Hong Kong), Croatia, Cyprus, Czech Re-
land, Italy, Latvia, Lithuania, Luxembourg, public, Denmark, Estonia, Finland, France,
Netherlands, Norway, Poland, Portugal, Germany, Greece, Hungary, Iceland, India,
Romania, Slovakia, Slovenia, Spain, Swe- Ireland, Israel, Italy, Japan, Latvia, Lithua-
den, Switzerland, Turkey, United Kingdom nia, Luxembourg, Malta, Mexico, Nether-
lands, Norway, Poland, Portugal, Roma-
nia, Russia, Slovakia, Slovenia, South Ko-
rea, Spain, Sweden, Switzerland, Turkey,
Ukraine, United Kingdom, United States

tutional risk on FDI into European transition economies, after controlling for other
factors. Wheeler and Mody (1992) use several risk factors, but did not find evidence
that these factors influence the location of US foreign affiliates.

3 Data
Our purpose is to explain inward FDI stocks from 45 source countries to 29 host
countries (see Table 1) for the 2006–2008 period. The literature has advocated the
use of FDI stocks relative to flows, since the former presents several advantages: they
are based on past accumulated flows, and hence they are less volatile; they are not
as influenced by specific year investments as flows are; they are the relevant decision
variable for a firm in the long term; and finally, they are a better measure of capital
ownership (Bénassy-Quéré et al., 2007). Moreover, since institutions are usually stable
over time, they are more likely to influence stocks rather than flows. Regressions
using stocks also attain a better fit, since one can use the logarithm of FDI as the
dependent variable due to the absence of negative values. FDI data—totaling 1144
observations—were collected from Eurostat database.4
We explain inward FDI according to an augmented gravity-type model, which
states that FDI depends on several types of variables: geographic, economic and
institutional. As for geographical factors, we include the physical distance between
host and source countries, which can be seen as a proxy for transaction costs (such
as transport costs, communication costs, and cultural and language differences), and
4
The data are described in detail in Appendix A.

6
a border dummy variable, which takes the value of 1 if the source and host countries
share a common border and 0 otherwise. A higher distance between source and host
countries is expected to have a negative impact on FDI, whereas a common border
between source and host countries is expected to increase FDI.
Our key economic variables are the host country’s GDP (a proxy for market size),
the GDP growth rate (a proxy for market growth) and per capita GDP (which reflects
a higher purchasing power or better socioeconomic conditions for the host country).
These 3 variables are expected to have a positive impact on FDI; however, it is not
uncommon in the literature to find per capita GDP having a negative sign in the
regressions, as a higher value is also associated with higher labor costs, which retracts
FDI (Bénassy-Quéré et al., 2007).5 The degree of openness—the share of imports
plus exports over GDP—is also expected to influence FDI, as it measures trade flows.
These variables were collected from the Eurostat database.
Our baseline model also includes the level of education, measured as the percent-
age of population aged 25 to 64 having completed secondary education. A priori,
education has an ambiguous effect on FDI—on the one hand, more education implies
higher labor productivity, but, on the other hand, it is associated to a higher wage
structure (Altomonte and Guagliano, 2003). We decided to use secondary education
instead of tertiary education in our analysis, since the correlation between the latter
and our institutional variables is much larger. We also experimented our model with
three different tax measures—the statutory tax rate, the effective marginal tax rate
(EMTR), and the effective average tax rate (EATR)—but our baseline model retains
only the EATR. The statutory tax rate is the relevant variable for companies seeking
to shift income towards low tax countries, whereas the EATR reflects the incentives
(such as investment tax credits and accelerated depreciation) that are granted to firms
when the investment occurs (Grubert and Mutti, 1991). The EMTR, on the other
hand, captures incentives to use new capital once the location choice has been made
(Devereux and Griffith, 1998). Not surprisingly (and confirming the literature on the
effects of corporate taxes on FDI), the EMTR came insignificant in all out regressions,
and was thus dropped in the final specification. As the statutory tax and the EATR
were highly correlated in our database, we decided to keep the EATR as representing
the effective tax rate paid by corporations, thus dropping also the statutory tax rate.
In fact, the EATR should be the relevant decision variable for multinationals seeking
to invest abroad (Devereux and Griffith, 1998). Education was taken from the Euro-
stat database, whereas effective tax rates were kindly provided by Michael Overesch
(see Overesch and Rincke, 2009). All variables are 2006–2008 averages, in order to
5
In our data, the correlation between per capita GDP and labor costs is around 90%.

7
smooth extreme events. The exception is the EATR, which is for 2006.
Our first set of institutional variables is based on the IEF, computed by the Her-
itage Foundation.6 The data collected concerns the year 2007, which covers the second
half of 2005 and the first half of 2006. As the Heritage Foundation puts on their web-
site, economic freedom is the fundamental right of every citizen to control his or her
own labor and property. “In a free society, individuals are free to make their own
production and consumption decisions, protected and unconstrained by the state.”
Thus, it is not surprising that societies that have better scores in this index are able
to attract more FDI, since they offer higher levels of protection to investors, lower tax
burdens, less restrictive regulations, less bureaucracy and less corruption. The IEF
is composed by 10 different components: business freedom, trade freedom, fiscal free-
dom, government freedom, monetary freedom, investment freedom, financial freedom,
property freedom, corruption freedom, and labor freedom. Each of these indexes was
rescaled to the 0–10 range, with higher scores meaning better performances. Our sec-
ond set of institutional variables is based on the political risk rating from the ICRG.
This rating assesses the political risk of the host country, and comprises 12 indicators:
government stability, socioeconomic conditions, investment profile, internal conflict,
external conflict, corruption, military in politics, religion in politics, law and order,
ethnic tensions, democratic accountability and bureaucracy quality. All variables were
converted to the 0–10 scale to ease comparisons. The data collected is for 2006.
Finally, our last set of institutional variables assesses the ease of doing business
in the host country. The DB database complements the more generic information on
business regulations that is reported by the IEF, by measuring the cost of starting,
operating, and closing a business, for a medium-sized firm in a given country.7 The DB
database reports 33 variables, covering 9 different areas—starting a business, dealing
with construction permits, registering property, getting credit, protecting investors,
paying taxes, trading across borders, enforcing contracts and closing a business. For
convenience, each of the 33 variables were converted to indexes, according to the
min-max standardization method (see the appendix). To ease interpretations, this
conversion was made such that higher values always mean better performances. The
resulting indexes were then summarized into these 9 categories. The data collected
respects the 2007 report, which addresses business regulations as of June 1, 2006.
6
The data for the IEF can be found at www.heritage.org/index.
7
The Doing Business report is a co-publication of the World Bank and the International Finance
Corporation, and the data is available at www.doingbusiness.org.

8
4 Econometric approach
The gravity model was first developed in the context of international trade (see,
for instance, Eaton and Tamura, 1995), but it has also been successfully applied to
explain bilateral FDI (e.g. Wei, 2000a,b). In its simplest formulation, the gravity
model states that bilateral FDI depends positively on the economic size of the source
and host countries and negatively on the distance between them. Here, we use an
augmented version of the gravity model to take into account other economic and
institutional factors that affect FDI. Denoting by j the source country and by i the
host country, we estimate the following augmented gravity-type model

log(FDIij ) = αcj + β1 DISTANCEij + β2 ECOi + β3 INSTi + εij (1)

Here, FDIij is the inward FDI stock from country j to country i; DISTANCEij is a
vector composed by the physical distance between country j and country i and the
border dummy variable; ECOi is a vector containing the host country economic vari-
ables, namely GDP, per capita GDP, GDP growth, the degree of openness, education
and the EART; and finally INSTi is a vector of institutional variables for the host
country. Besides FDIij , the distance between source and host countries, GDP and
per capita GDP will enter (1) in logarithmic form, which helps making the error term
homoskedastic.8 Furthermore, a double-log specification displays the best fit to the
data, consistently delivering good values for the R2 and more precise estimates as
compared to alternative specifications (e.g. Stein and Daude, 2007). We implement a
quasi-fixed effects model, i.e., we include source country dummies, represented in (1)
by the vector cj . These dummy variables are meant to capture all specific character-
istics of the source country that are relevant to the size of outward FDI, such as the
level of GDP, the level of development or the institutional framework. Host country
dummies are not included, since doing so would eliminate the possibility of estimating
all the coefficients of the model, much in the same way as time invariant regressors
cannot be estimated in a panel equation using a fixed-effects estimator. Finally, εij is
an i.i.d. error term which is assumed normally distributed.
To estimate the double-log model in equation (1) by OLS, all zero-FDI observations
have to be dropped, since the logarithm of zero is not defined. In our case, this
corresponds to 202 observations—about 17.5% of our sample. This obviously results
in a censored-sample problem, which can lead to inconsistency. A common way to
retain these zero observations is to use a Tobit model (e.g. Stein and Daude, 2007,
8
Nonetheless, heteroskedasticity consistent standard errors will be reported.

9
Gao, 2005). This approach can by justified by considering that stocks below a certain
threshold are incorrectly recorded as zeros, or that the desired level of investment is
positive, but the presence of fixed costs of investing abroad leads to observed zero-
FDI values when the desired investment is below a certain threshold.9 Hence, besides
estimating (1) by OLS, we also estimate a Tobit model where the assumed threshold
is −1.1.10
Institutional indicators are highly correlated among them, which may originate
problems of near multi-collinearity if several of these variables are simultaneously
included in the regressions. In this case, the resulting OLS estimator has a low
probability of being close to its true value, due to variance inflation (Hwang and
Nettleton, 2003). We tackle the problems caused by correlated institutional variables
through three distinct approaches. First, for each institutional database—IEF, ICRG,
and DB—we run a Principal Component Analysis (PCA), in order to summarize the
information in a smaller set of variables (components).11 The resulting components
can, in general, be associated with an identifiable institutional area, although the
resulting aggregation is usually too broad to extract conclusions about the effects of
any particular institutional indicator on inward FDI. Besides reducing the dimension
of each database, the resulting score vectors are orthogonal, thus diminishing the
correlations between institutional factors in our sample. The PCA is applied to all
countries within each institutional database (and not only to the host countries used
herein), and is followed by a varimax (orthogonal) rotation. These new variables are
then used in (1) to measure the effects of institutions on FDI.
To assess the robustness of these results, we also reduce the dimension of our
database through a direct aggregation of variables. In this case, we use the informa-
tion of the rotated factor loadings matrix to identify those indicators that are highly
correlated among them, and which can therefore be aggregated into a new variable.
The aggregation was done by taking the simple average, and the resulting variables
can be interpreted as factor-based scores.
The main disadvantage of these approaches is that the resulting aggregation is
too broad, because the correlations between several variables are usually high. For
9
Desired FDI (the latent variable) is equal to the realized FDI for values above a certain threshold,
but is not observed for values below that threshold. Some authors have proposed to use log(1+FDIij )
or log(a + FDIij ) (where a is a parameter to be estimated) instead of log FDIij as the dependent
variable in order to estimate (1) while retaining zero-FDI observations. However, this approach is
completely add hoc, and the results depend on the measurement unit.
10
The minimum value of the average of inward FDI stocks for the 2006-2008 period is 1/3 million
euros, and log(1/3) ≈ −1.1.
11
This was preceded by the KMO and Bartlett’s test of sphericity, which indicate whether the
PCA is appropriate or not.

10
instance, the effect of bureaucracy on FDI cannot be disentangled from the effect of
corruption, or from the effect of the costs of starting or closing a business, since these
variables are highly correlated. To tackle this issue, we estimate the model in (1) by
adding each institutional variable successively. This approach is widely followed in
the literature (e.g. Walsh and Yu, 2010, Chakrabarti, 2001), and to our knowledge, it
is the only way to evaluate the effect of individual institutions of FDI while avoiding
the problems caused by variance inflation. This approach should be interpreted as an
attempt to explore possible correlations between institutional indicators and inward
FDI, rather than to explore any link of causality, and our results should be interpreted
taking this setback into account.

5 Results
5.1 Economic determinants of FDI
Table 2 reports the results for the baseline model. No institutional variable is consid-
ered here. The coefficients obtained with OLS, reported in columns (1) and (3), do
not differ substantially from those in columns (2) and (4), for the Tobit model. This
suggests that the censored-sample problem is not serious in our sample. According to
Table 2, inward FDI stocks are characterized by strong border effects: the investment
of a country in its neighbor is about 105–116% (e0.72 − 1 ≈ 1.05 and e0.77 − 1 ≈ 1.16)12
higher as compared to the investment in another country with similar characteristics,
but with which the source country does not share a common border. Distance is also
a key determinant of inward FDI, as an increase of 1% in the number of kilometers
between source and host countries reduces FDI between 1.23% and 1.32%. GDP
presents a statistically significant impact on FDI, giving support to the market size
hypothesis.13 The effect of per capita GDP on FDI is significant, but negative. Recall
that a higher per capita GDP is associated with higher standards of living and better
infrastructures, but also with higher labor costs. The sign of the coefficient suggests
that this latter effect dominates the former. GDP growth has also a negative impact
on FDI. This impact is negligible if one does not account for education and taxes, but
it is statistically significant at 5% once we control for the effects of these variables on
inward FDI. As we show later, the absence of institutional determinants from these
regressions can partially explain the negative sign of the coefficient, since better insti-
12
In this article, we use this formula to compute all marginal effects when the regressor is not in
logarithmic form.
13
Unless specified otherwise, we use a 5% significance level.

11
Table 2: Baseline regression results.
(1) (2) (3) (4)
OLS Tobit OLS Tobit
border 0.7200∗∗∗ 0.7345∗∗∗ 0.7561∗∗∗ 0.7698∗∗∗
(0.2241) (0.2233) (0.2244) (0.2232)
log distance -1.2954∗∗∗ -1.3223∗∗∗ -1.2395∗∗∗ -1.2679∗∗∗
(0.1184) (0.1182) (0.1233) (0.1227)
log gdp 0.9944∗∗∗ 1.0200∗∗∗ 1.0749∗∗∗ 1.1073∗∗∗
(0.0591) (0.0602) (0.0697) (0.0709)
log gdp per capita -0.3189∗∗∗ -0.3332∗∗∗ -0.3358∗∗∗ -0.3422∗∗∗
(0.1205) (0.1201) (0.1286) (0.1276)
gdp growth -0.0583 -0.0594 -0.1129∗∗ -0.1145∗∗
(0.0398) (0.0400) (0.0465) (0.0463)
openness 0.0093∗∗∗ 0.0096∗∗∗ 0.0085∗∗∗ 0.0087∗∗∗
(0.0017) (0.0017) (0.0016) (0.0016)
secondary education 0.0092∗∗ 0.0092∗∗
(0.0044) (0.0044)
eatr -0.0280∗∗ -0.0305∗∗
(0.0138) (0.0137)
adjusted R2 0.72 n.a. 0.72 n.a.
White-robust standard errors in parenthesis. ∗ , ∗∗ and ∗∗∗ represent rejections at 10, 5 and 1 percent significance
levels, respectively.

tutions affect growth. GDP growth may also be endogenous, since it has been shown
that greater amounts of FDI can have positive repercussions on economic growth (e.g.
Herzer, 2008, 2010, Borensztein et al., 1998). However, it is not our purpose to tackle
this issue here. Finally, openness is also statistically significant and has the expected
sign: an increase in this variable by 1 percentage point (p.p.) increases FDI around
0.9%. Columns (3) and (4) add education and the EATR to the regression. The
coefficient for education states that an increase of 1 p.p. in the population aged 25 to
64 having completed secondary education leads to an increase in FDI below 1%—an
impact that, although small, suggests that education may affect FDI. However, one
must take into account that part of this effect may be related with institutions as well,
since countries with higher levels of education have also better institutions. Finally,
the EATR is statistically significant and has the expected sign: a decrease in this
variable by 1 p.p. increases FDI around 2.8–3.1%.

5.2 Institutional determinants of FDI

We now add the institutional dimension to the model. We undertake 3 different
approaches. The first is based on the PCA, whereas the second consists in a factor-
based scores approach. The objective of these two approaches is to aggregate the
information of several institutional indicators in a few set of variables, which can

12
therefore be simultaneously included in the regressions. The last approach consists in
including each institutional indicator successively in the regressions, in order to assess
the effects of specific institutional indicators in FDI while avoiding variance inflation
issues.

A principal components analysis

First, we use the PCA to summarize the indicators from the IEF, from the political
risk index of the ICRG, and from the DB database, in a set of few variables, which
can clearly be associated to an identifiable institutional area. We then run the regres-
sions using these newly created variables to represent the institutional framework of
a country.
For the IEF, the standard eigenvalue-based criteria, with a cut-off value of 1,
identified 2 components, explaining almost 68% of total variance. The rotated factor
loading matrix associates the first component score (hereinafter denominated as “firms
freedom”) to all the elements that, in some way, influence the firms profitability and
the ease of doing business. It includes the following indicators: business freedom,
trade freedom, monetary freedom, investment freedom, financial freedom, property
rights, freedom from corruption, and labor freedom. All these variables displayed
a factor loading above 0.62 in this component, whereas the correlation between the
remaining variables and this component was below 0.2. The second component score
(which we call “public sector freedom”) measures the public sector effects on economic
freedom—namely fiscal freedom and government freedom.
When applied to the 12 political risk variables from the ICRG, the PCA identifies
3 components, which together explain more than 72% of total variance. The first
component is interpreted as the firms political risk, as it is related with the politi-
cal risk factors that may directly influence a firms performance. The rotated factor
loadings associate this component with socioeconomic conditions, investment profile,
corruption, law and order, bureaucracy quality and military in politics. It may seem
surprising that the variable military in politics loads in this component, but the fact
is that a country with a military regime may have an uneasy environment for for-
eign businesses, and is more prone to corruption. The second component represents
conflicts and tensions, as it is highly correlated with the variables internal conflict,
external conflict, religious tensions and ethnic tensions. The last component stands
for democratic responsiveness, and comprises government stability and democratic
accountability. In this component, government stability loads with a negative value,
and hence higher values are associated with greater democratic accountability, but

13
Table 3: Regression results—Institutional determinants under a PCA approach.
(1) (2) (3) (4) (5) (6)
OLS Tobit OLS Tobit OLS Tobit
border 0.7615∗∗∗ 0.7796∗∗∗ 0.7421∗∗∗ 0.7594∗∗∗ 0.7430∗∗∗ 0.7603∗∗∗
(0.2255) (0.2232) (0.2263) (0.2237) (0.2266) (0.2242)
log distance -1.3102∗∗∗ -1.3294∗∗∗ -1.2971∗∗∗ -1.3172∗∗∗ -1.3153∗∗∗ -1.3359∗∗∗
(0.1277) (0.1261) (0.1279) (0.1260) (0.1275) (0.1259)
log gdp 1.0444∗∗∗ 1.0692∗∗∗ 0.9485∗∗∗ 0.9748∗∗∗ 1.0160∗∗∗ 1.0413∗∗∗
(0.0755) (0.0754) (0.0760) (0.0761) (0.0755) (0.0755)
log gdp per capita -0.8818∗∗∗ -0.8883∗∗∗ -1.1844∗∗∗ -1.1966∗∗∗ -0.7478∗∗∗ -0.7557∗∗∗
(0.1762) (0.1746) (0.2309) (0.2302) (0.1671) (0.1660)
gdp growth -0.0331 -0.0350 -0.1390∗∗∗ -0.1416∗∗∗ -0.0957∗ -0.0976∗∗
(0.0515) (0.0512) (0.0508) (0.0506) (0.0496) (0.0494)
openness 0.0054∗∗∗ 0.0056∗∗∗ 0.0062∗∗∗ 0.0064∗∗∗ 0.0086∗∗∗ 0.0088∗∗∗
(0.0017) (0.0017) (0.0017) (0.0017) (0.0016) (0.0016)
secondary education 0.0003 0.0002 0.0036 0.0034 0.0068 0.0067
(0.0051) (0.0052) (0.0050) (0.0050) (0.0048) (0.0049)
eatr -0.0177 -0.0197 -0.0013 -0.0036 -0.0106 -0.0126
(0.0138) (0.0136) (0.0147) (0.0145) (0.0143) (0.0142)
Institutional components
IEF—firms freedom 0.8209∗∗∗ 0.8241∗∗∗
(0.1532) (0.1536)
IEF—public sector freedom -0.1897∗ -0.1909∗∗ 0.1916∗ 0.1904∗ -0.0296 -0.0294
(0.0974) (0.0963) (0.1005) (0.1003) (0.0907) (0.0899)
ICRG—firms political risk 1.1942∗∗∗ 1.2021∗∗∗
(0.2515) (0.2514)
ICRG—conflicts and tensions -0.0750 -0.0838 0.0822 0.0779 -0.1556 -0.1642∗
(0.1002) (0.0993) (0.1234) (0.1231) (0.0956) (0.0948)
ICRG—democratic responsiveness -0.0420 -0.0260 0.2332 0.2503 0.0203 0.0360
(0.1637) (0.1632) (0.1740) (0.1747) (0.1648) (0.1647)
DB—overall 0.5646∗∗∗ 0.5682∗∗∗
(0.1091) (0.1103)
adjusted R2 0.73 n.a. 0.73 n.a. 0.73 n.a.
White-robust standard errors in parenthesis. ∗ , ∗∗ and ∗∗∗ represent rejections at 10, 5 and 1 percent significance
levels, respectively.

with lower government stability. This occurs because a higher government stability
is associated with a greater government’s ability to stay in office, and sometimes this
is achieved at the expense of a lower democratic accountability (e.g. one party states
or autocracies).
The standard eigenvalue-based criteria applied to the 9 areas of doing business
identified 2 factors; however factor loadings were perplexing and did not yield any
interesting conclusion. We opted instead to extract only one factor from the DB
database, which is interpreted as representing an overall measure of doing business.
This component represents 45% of total variance, and is positively correlated with all
the 33 index variables.
The results are presented in Table 3. As firms freedom, firms political risk, and the

14
doing business component are highly correlated, and shared similar indicators, we did
not include them simultaneously in the regressions. According to columns (1) and (2),
a higher firms freedom has a statistically significant and positive impact on inward
FDI at a 1% significance level. Columns (3) to (6) confirm that political risk and the
ease of doing business are also key determinants of inward FDI. Public sector freedom
has an ambiguous impact on FDI, depending on which control—firms freedom, firms
political risk and the ease of doing business—is included. Fiscal freedom assesses the
fiscal burden of a society, with more freedom being associated with lower taxes. As
it includes the top tax rate on corporate income, ceteris paribus, one should expect
higher values in fiscal freedom to be associated with more FDI. Government freedom
measures the level of government expenditures as a percentage of GDP, with more
freedom being associated with lower expenditures. It is not clear whether this should
attract or repel FDI, since higher public expenditures may be associated, on the one
hand, with better socioeconomic conditions, higher development, better infrastruc-
tures, or greater incentives for FDI, but, on the other hand, with a higher future fiscal
burden and fiscal uncertainty, or to a lower efficiency in the usage of public resources.
This may help explaining why the results were unclear. As expected, conflicts and
tensions, per se, do not have a statistically significant impact on inward FDI. Hence,
multinational firms do not seem to care directly about conflicts and tensions in the
host country, as long as their investments are protected by the state, and the eco-
nomic, legal and democratic environment is conductive to doing business. A similar
reasoning can be applied to democratic responsiveness.
It is also worthwhile to note that, in columns (1) and (2), the coefficient for GDP
growth is insignificant, whereas in columns (3) to (6), as in the baseline regression,
it is significant and negative. One possible interpretation for this result is that low
growth countries have a higher level of development and hence better institutions,
and the variable firms freedom provides a better control for those institutions that
influence the performance of firms. Once we control for this effect, economic growth
becomes irrelevant to explain inward FDI. Obviously, the endogeneity problem men-
tioned earlier may also affect the level of significance. Education was only significant
in the baseline regression insofar as it was capturing the effects of institutions on in-
ward FDI. Once institutions are incorporated in the model, the effect of education on
FDI becomes irrelevant. The coefficient for the EATR has also become insignificant
in the regressions in Table 3, suggesting that corporate taxes have a secondary role in
FDI attractiveness as compared to institutions.

15
A factor-based scores approach
We now use the information from the rotated factor loadings matrix to aggregate the
variables in an economically meaningful way. This robustness exercise has two ad-
vantages. First, it allows us to protect against the fact that factor scores representing
a given component always have some residual correlations with other components,
which might affect the results. Second, contrary to factor scores, the vectors com-
puted here are interpretable, as the variables are measured in indexes. A potential
disadvantage of this approach is that the weights are not optimally computed. The
new variables were aggregated by taking the simple average across the indicators that
the PCA identified as loading into the same component. The exception is democratic
responsiveness, which does not consider government stability. In fact, as this variable
is negatively associated with democratic accountability by construction (the indicator
loaded in the component with a negative sign), it makes no sense to take the simple
average between them to create a factor of democratic responsiveness.
The results are presented in Table 4. Firms freedom, firms political risk and the
ease of doing business have a positive and statistically significant impact on inward
FDI, confirming our previous results (again, these components were not considered
simultaneously in the regressions). For instance, an increase in 1 point in the firms
freedom index (in a 0-10 scale) increases FDI around 80%. The effect is even higher
for the ease of doing business, as a 1 point increase in this index more than doubles
inward FDI. The effects of public sector freedom and conflicts and tensions on inward
FDI are inconclusive, just as before, whereas democratic responsiveness (democratic
accountability) seems to have a positive impact on inward FDI. This analysis confirms
that countries with better institutions—more specifically with a better economic and
business environment and lower bureaucratic load—are able to attract larger amounts
of FDI, and the effects are significant and important.

Institutional analysis—a breakdown

The previous analysis allowed us to ascertain the effects of institutions in inward FDI
at a broad level, but it was silent about the effects of specific institutions. The main
difficulty in assessing the effects of specific institutions on FDI is that institutional
indicators are highly correlated among themselves, for which the regressions which
include several of these variables simultaneously can perform very poorly. To tackle
this issue, we now re-estimate (1) by successively adding each institutional variable to
the model. Hence, we are not interested in determining causality here, but just to find
patterns of correlation between institutional indicators and long-run FDI. We omit the

16
Table 4: Regression results—Institutional determinants under a factor-based scores
approach.
(1) (2) (3) (4) (5) (6)
OLS Tobit OLS Tobit OLS Tobit
border 0.7554∗∗∗ 0.7729∗∗∗ 0.7507∗∗∗ 0.7679∗∗∗ 0.7445∗∗∗ 0.7614∗∗∗
(0.2267) (0.2242) (0.2271) (0.2245) (0.2274) (0.2249)
log distance -1.3413∗∗∗ -1.3598∗∗∗ -1.2993∗∗∗ -1.3192∗∗∗ -1.3388∗∗∗ -1.3586∗∗∗
(0.1280) (0.1261) (0.1280) (0.1261) (0.1275) (0.1257)
log gdp 0.9174∗∗∗ 0.9440∗∗∗ 0.8774∗∗∗ 0.9047∗∗∗ 0.9000∗∗∗ 0.9272∗∗∗
(0.0812) (0.0810) (0.0814) (0.0813) (0.0808) (0.0807)
log gdp per capita -0.9352∗∗∗ -0.9498∗∗∗ -0.9830∗∗∗ -0.9922∗∗∗ -0.7393∗∗∗ -0.7496∗∗∗
(0.1764) (0.1754) (0.2215) (0.2210) (0.1638) (0.1627)
gdp growth -0.0488 -0.0498 -0.1339∗∗ -0.1354∗∗ -0.1036∗ -0.1047∗∗
(0.0549) (0.0543) (0.0537) (0.0532) (0.0536) (0.0531)
openness 0.0042∗∗ 0.0044∗∗ 0.0060∗∗∗ 0.0063∗∗∗ 0.0076∗∗∗ 0.0079∗∗∗
(0.0017) (0.0017) (0.0017) (0.0017) (0.0016) (0.0016)
secondary education 0.0003 0.0003 0.0065 0.0065 0.0074 0.0073
(0.0049) (0.0048) (0.0048) (0.0047) (0.0047) (0.0046)
eatr -0.0056 -0.0076 -0.0030 -0.0054 -0.0040 -0.0061
(0.0145) (0.0142) (0.0149) (0.0147) (0.0149) (0.0147)
Institutional components
IEF—firms freedom 0.5861∗∗∗ 0.5929∗∗∗
(0.0997) (0.0998)
IEF—public sector freedom -0.0156 -0.0171 0.1004∗ 0.0969∗ -0.0396 -0.0412
(0.0492) (0.0488) (0.0515) (0.0515) (0.0515) (0.0509)
ICRG—firms political risk 0.5297∗∗∗ 0.5297∗∗∗
(0.1312) (0.1306)
ICRG—conflicts and tensions -0.0525 -0.0588 -0.2846∗∗∗ -0.2903∗∗∗ -0.1402∗ -0.1470∗
(0.0804) (0.0796) (0.0815) (0.0806) (0.0774) (0.0557)
ICRG—democratic responsiveness 0.3206∗∗ 0.3293∗∗∗ 0.2395∗ 0.2477∗ 0.2932∗∗ 0.2995∗∗
(0.1263) (0.1248) (0.1317) (0.1299) (0.1271) (0.1257)
DB—overall 0.7127∗∗∗ 0.7175∗∗∗
(0.1326) (0.1331)
adjusted R2 0.73 n.a. 0.73 n.a. 0.73 n.a.
White-robust standard errors in parenthesis. ∗ , ∗∗ and ∗∗∗ represent rejections at 10, 5 and 1 percent significance
levels, respectively.

indicators related with conflicts and tensions and to public sector freedom from the
analysis, since our previous results suggest that these variables have a dubious impact
on FDI. This also accords with our intuition that the key institutional determinants
of inward FDI are those which are directly related with the restrictions to economic
activity and to doing business.
The results are presented in Table 5. To save space, we omit the coefficients for
the control variables, and show only the coefficients for the institutional indicators.
The variables are ordered according to their contribution in explaining the variance of
the dependent variable (R2 ) in the OLS estimation. Recall that all indicators range
from 0 to 10, so that the coefficients must be interpreted as the impact on FDI from a

17
Table 5: Regression results—a breakdown.
(1) st. dev. (2) st. dev.
OLS OLS R2 Tobit Tobit
IEF variables
financial freedom 0.2439∗∗∗ 0.0424 0.7314 0.2484∗∗∗ 0.0424
freedom from corruption 0.2752∗∗∗ 0.0505 0.7300 0.2765∗∗∗ 0.0505
business freedom 0.3128∗∗∗ 0.0622 0.7287 0.3182∗∗∗ 0.0632
labor freedom 0.1514∗∗∗ 0.0420 0.7254 0.1575∗∗∗ 0.0422
property rights 0.1409∗∗∗ 0.0458 0.7243 0.1415∗∗∗ 0.0459
investment freedom 0.1488∗∗∗ 0.0491 0.7238 0.1427∗∗∗ 0.0486
trade freedom -0.4289∗∗∗ 0.1631 0.7237 -0.4516∗∗∗ 0.1648
monetary freedom 0.2306 0.1718 0.7219 0.2281 0.1714
ICRG variables
corruption 0.1635∗∗∗ 0.0430 0.7258 0.1648∗∗∗ 0.0427
law and order 0.2567∗∗∗ 0.0690 0.7253 0.2619∗∗∗ 0.0685
socioeconomic conditions 0.2603∗∗∗ 0.0934 0.7239 0.2639∗∗∗ 0.0942
investment profile 0.2457∗∗ 0.0973 0.7236 0.2478∗∗ 0.0989
democratic accountability 0.2738∗∗ 0.1262 0.7231 0.2818∗∗ 0.1253
military in politics -0.0570 0.0788 0.7216 -0.0610 0.0794
bureaucracy quality -0.0010 0.0525 0.7214 -0.0063 0.0529
doing business variables
starting a business 0.7381∗∗∗ 0.1501 0.7291 0.7567∗∗∗ 0.1509
protecting investors 0.2119∗∗∗ 0.0478 0.7280 0.2109∗∗∗ 0.0475
getting credit 0.2320∗∗∗ 0.0541 0.7271 0.2322∗∗∗ 0.0550
construction permits 0.4254∗∗∗ 0.1058 0.7255 0.4369∗∗∗ 0.1073
closing a business 0.1854∗∗∗ 0.0502 0.7250 0.1833∗∗∗ 0.0500
trading across borders 0.4587∗∗∗ 0.1374 0.7251 0.4776∗∗∗ 0.1413
enforcing contracts 0.1856∗∗∗ 0.0698 0.7236 0.1868∗∗∗ 0.0696
registering property 0.0671 0.0590 0.7218 0.0657 0.0594
paying taxes 0.0962 0.1404 0.7215 0.0804 0.1408
White-robust standard errors in parenthesis. ∗ , ∗∗ and∗∗∗ represent rejections at 10, 5 and 1 percent significance
levels, respectively. Coefficients for control variables are omitted to save space.

change in 1/10 points of the range of the respective index. Higher values always mean
better performances. For the IEF, the results indicate that all variables, except trade
freedom, display the expected sign, and all of them, except monetary freedom, are
significant at 1%. The sign of the coefficient for trade freedom is somewhat surprising,
since one would except lower tariffs and less tariff barriers in the host country to
increase inbound FDI. To check if this result was due to an omitted variable bias, we
also experimented to augment the regression with other institutional factors from the
IEF, but the negative sign of the coefficient remained quite robust to these alternative
specifications.14 However, our database is mostly composed by EU countries, which
14
Only a few variables were considered at a time to avoid multi-collinearity issues.

18
have a common trade policy. This implies that the variability of the trade freedom
indicator in our sample is extremely low, so that our quasi-fixed effects estimator may
be highly inefficient in this case. The same holds for monetary freedom. This indicator
addresses price stability and the absence of price controls, characteristics which do
not vary substantially across the EU countries. The variability of other indicators is
much higher, and so this problem does not arise.
The most relevant factors affecting inbound FDI, in the sense that they explain
a greater fraction of the variance of the dependent variable, are financial freedom
and freedom from corruption. These indicators have a unitary impact on inward FDI
around 28% and 32% respectively. Financial freedom measures the independence of
financial institutions from state control, which contributes to more competition and to
a higher level of services available from financial intermediaries. Hence, more freedom
at this level means that more financial services are available for multinational firms
in the host country. Corruption introduces insecurity and uncertainty into economic
relationships, and increases the pecuniary and non-pecuniary costs of operating a
business. Thus, these indicators have a natural impact on inward FDI. Business
freedom and labor freedom—which also play an important role in explaining the
variance of the dependent variable—are directly linked to the potential performance
of firms, as well as to their costs of adjustment. The former measures the overall
burden and government efficiency associated to starting, operating, and closing a
business, whereas the latter provides a quantitative measure of labor regulations: the
higher the score, the more flexible is the labor market and the less expensive is for
multinational corporations to adjust their labor force. Hence, it is not surprising that
higher factor scores in these indexes are associated to higher volumes of FDI. In fact,
a one point increase in business freedom leads to an increase in FDI around 37%,
whereas the same score increase in labor freedom has an impact on inbound FDI
around 16–17%. Finally, property rights—which assesses the ability of individuals to
accumulate private property, the extent to which laws protect that property, and the
efficiency of the judiciary system to enforce those laws—and investment freedom—
which addresses the constraints on capital flows, both in and out specific activities
and across borders—also play a role in fostering inward FDI, although to a lesser
extent. According to Table 5, a unit increase in these indicators leads to an increase
in FDI around 15–16%.
The second set of results in Table 5 confirm the idea that low political risk, sup-
ported by good institutions, fosters inbound FDI in the long-run. Military in politics
and bureaucracy quality have a non-significant impact on inward FDI, as expected:
military in politics could only affect inward FDI insofar as it is correlated with cor-

19
ruption, but this effect does not seem to be present here; and bureaucracy quality
is irrelevant insofar as it only measures the extent to which administrative functions
are independent from the political sphere, but does not seem to capture the effects
of the bureaucratic burden on firms. From the variables of the ICRG, corruption is
the one which contributes the most to explain the variance of the dependent variable.
This reinforces the our previous conclusion, and suggests that corruption can indeed
impose an extra burden on businesses, by increasing the costs of doing business and
the risks of social unrest. The effects of the effectiveness, strength and impartial-
ity of the judicial system and of the popular observance of the law (law and order),
and of socioeconomic conditions are also significant at 1%, and have the predicted
sign: a 1 point increase in any of these variables fosters inward FDI around 29–30%.
Democratic accountability, which measures the ability of the government to be held
accountable for its actions, is also significant, although only at 5%. Finally, the effect
of the investment profile indicator goes in the same direction of the effect of invest-
ment freedom, and confirms the idea that the risks related with expropriations, the
restrictions on repatriation of profits, and payment delays, can deter foreign invest-
ments. All in all, these results indicate that a stable and well-functioning democracy
can boost inward FDI.
All the coefficients for the 9 areas of doing business are positive. From these, 2 are
insignificant: registering property and paying taxes. Our intuition also suggests that
these factors should only influence inward FDI at the margin. The remaining variables
are significant at 1%. The most important is starting a business, whose unitary
increase fosters FDI in more than 100%. The importance of this indicator was already
identified by the business freedom coefficient, although the effect here is much larger.
The difference in magnitudes between both variables is most probably explained by
the fact that business freedom includes other variables besides those considered in
starting a business, with a lower impact in inward FDI. The formulas used to compute
both indexes are different also. The strength of investor protection, measured by the
protecting investors index, and the credit information registries and the effectiveness
of collateral and bankruptcy laws in facilitating lending, measured by the getting
credit index, also provide an important contribute to explain the variability of the
dependent variable, and have an important impact on FDI: a one point increase in each
of these variables leads to an increase in FDI around 23–26%. From the remaining,
the licensing procedures for some activities (dealing with construction permits) and
the necessary procedural requirements for exporting and importing (trading across
borders) display also a significant effect on inward FDI, over 50% for a unit increase.
Finally, better performances in the time, cost and number of procedures required to

20
enforce a contract, and in the time, cost and recovery rate of closing a business, also
increase the amount of inward FDI, although to a lesser extent.
The analysis in this section clearly confirms that institutional factors are an im-
portant driving force of foreign investments. The number of procedures, the costs and
the time require to start and operate a business, the level of corruption, the amount
of financial services provided in the host country, the extent of investors protection,
the effectiveness of collateral and bankruptcy laws in facilitating lending, and labor
market flexibility are the issues which most contribute to explain inward FDI. How-
ever, the number of procedures, the costs and the time require to start a business, the
procedures, time and costs related with licensing procedures, and the necessary re-
quirements for exporting and importing are the areas which display the highest effects
on inward FDI.

6 Measuring institutional reform in Portugal

If one seeks to understand the impact on Portugal’s inbound FDI originating from
an institutional reform that approximates Portuguese institutions to the EU level,
the analysis from the previous section provides only an incomplete picture. Firstly,
because it does not incorporate the amount of institutional reform needed to achieve
the EU level. And secondly, because it does not address the effort required to achieve
that level. In this section we evaluate the effects of an institutional reform in Portugal.
We follow closely Tavares (2004), who have proposed 3 measures of institutional reform
to assess the benefits of reforming Portuguese institutions to the level of the EU on
Portuguese economic growth.
For each institutional indicator for which Portugal is lagging behind the EU, we
evaluate the impact on Portugal’s inward FDI of reforming that institution to the EU
average, the required reform effort and the efficiency of the reform. The analysis is
conducted versus the EU-27 and the EU-15.15 The impact of reforming institution
k to the EU level is given by the exponential of the respective coefficient for that
institution (computed in the previous section) multiplied by the difference between
the institutional index for the EU and for Portugal. That is


Impact on FDIk = exp β3,k (INSTEU,k − INSTP,k ) (2)
where INSTi,k denotes the institutional index of institution k in country i (where EU
15
For the EU-27, the analysis actually comprises only 25 countries, since Cyprus and Malta are
not included in our database.

21
stands for the EU average and P for Portugal) and β3,k is the respective coefficient.16
Obviously, the higher the value of (2), the more promising is the reform in that
area, either because it has a large impact on inbound FDI, or because the Portuguese
institutional index is substantially below that of the EU. For this reason, this measure
completely abstracts from the “cost of reform,” i.e., from the required effort to bring
the Portuguese institutional index closer to that of the EU. To measure this, Tavares
(2004) have proposed the following indicator, here adapted to our framework

INSTEU,k − INSTP,k
Required reform effortk = (3)
INSTP,k

Equation (3) measures the relative distance of the Portuguese institutional index
relative to the EU average, i.e., the required institutional change that Portugal needs
to achieve the EU level, relative to its current position. A higher value means that
achieving the EU average requires a higher percentage change in the institutional
indicator, and thus more effort has to be put on the reform.
Finally, the third measure of institutional reform evaluates the efficiency of the
reform, i.e., the impact on FDI for each unit of effort put in the reform. It is computed
as the ratio of (2) over (3)

Impact on FDIk
Efficiency of reformk = (4)
Required reform effortk

A value of 1 indicates a 100% increase in inbound FDI for each reform effort of 100%.
Hence, the highest the value of (4), the more promising is the reform in that area,
and the highest is the increase in FDI for each unit of effort put in the reform.
Tables 6 and 7 present an evaluation of the reform potential for our selected
institutional factors. We do this exercise only for the model estimated by OLS, since
the regression coefficients are similar to those from the Tobit model. Furthermore,
since it only makes sense to address the reform potential in areas where Portugal is
lagging behind the EU level, we do not present the values when the opposite situation
occurs, i.e. when Portugal has better institutions than the EU. Finally, note that
any area which has not a statistically significant impact in FDI should also not be
considered for reform.
In Table 6 we observe that two of the most promising areas for reform are financial
freedom and labor freedom, since these are expected to have the largest impact on
16
In our computations, we take into account that a change in the Portuguese institutional index
also changes EU average institutional index.

22
Table 6: Reforming institutional factors. Impact on FDI, required reform effort, and
efficiency of reform versus the EU-27.
(3) (5) (6) (7)
(1) (2) (4)
(2) − (1) e(4)(3) − 1 |(3)/(1)| (5)/(6)
impact required
index index efficiency
difference coefficient on FDI reform
Portugal EU-27 of reform
(%) effort
Institutional variables
firms freedom∗∗∗ 6.78 7.34 0.57 0.5861 39.29 8.34 4.71
firms political risk∗∗∗ 8.20 7.96 -0.24 0.5297
doing business∗∗∗ 7.84 7.68 -0.16 0.7127
IEF variables
financial freedom∗∗∗ 5.00 7.08 2.08 0.2439 66.22 41.67 1.59
corruption freedom∗∗∗ 6.50 6.46 -0.04 0.2752
business freedom∗∗∗ 7.86 8.00 0.14 0.3128 4.53 1.80 2.51
labor freedom∗∗∗ 4.15 6.34 2.19 0.1514 39.38 52.84 0.75
property rights∗∗∗ 7.00 6.96 -0.04 0.1409
investment freedom∗∗∗ 7.00 7.29 0.29 0.1488 4.44 4.17 1.06
trade freedom∗∗∗ 8.66 8.50 -0.16 -0.4289
monetary freedom 8.04 8.10 0.06 0.2306 1.33 0.72 1.87
ICRG variables
corruption∗∗∗ 6.67 5.94 -0.72 0.1635
law and order∗∗∗ 8.33 8.24 -0.10 0.2567
socioec. conditions∗∗∗ 6.70 6.60 -0.10 0.2603
investment profile∗∗ 10.00 9.53 -0.47 0.2457
democratic account.∗∗ 10.00 9.62 -0.38 0.2738
military in politics 10.00 9.40 -0.60 -0.0570
bureaucracy quality 7.50 8.02 0.52 -0.0010
doing business variables
starting a business∗∗∗ 9.12 9.20 0.08 0.7381 5.94 0.86 6.93
protecting investors∗∗∗ 6.08 5.57 -0.51 0.2119
getting credit∗∗∗ 5.56 4.66 -0.90 0.2320
construction permits∗∗∗ 8.34 8.94 0.60 0.4254 28.97 7.17 4.04
closing a business∗∗∗ 8.45 7.57 -0.88 0.1854
trading across borders∗∗∗ 8.84 8.69 -0.15 0.4587
enforcing contracts∗∗∗ 7.49 7.73 0.24 0.1856 4.64 3.26 1.42
registering property 7.70 8.02 0.32 0.0671 2.17 4.16 0.52
paying taxes 9.01 8.80 -0.21 0.0962
∗ , ∗∗ and ∗∗∗ represent the variables which are significant at 10, 5 and 1 percent significance levels, respectively. The

reform measures are only computed for those variables in which Portugal has an inferior performance relative to the
European Union. The index for the EU-27 does not consider Portugal.

Portuguese inbound FDI. However, if the effects of an institutional reform are weighted
versus the required reform effort, reforming the financial sector comes at a much lower
cost per unit of impact on FDI. Business freedom also displays a high reform efficiency,
but the overall impact on FDI is small, and hence it should not be considered a top
priority for reform. By the same token, reforming investment freedom does not bring
relevant gains as compared to the cost of reform. As for the DB variables, the licensing
procedures for some activities turns out to be a area with a high reform efficiency,
and the impact on FDI—about 29%—is quite relevant. The remaining areas of doing

23
business where Portugal is lagging behind the EU-27—starting a business, enforcing
contracts and registering property—do not have a relevant impact on FDI. Portugal
is above the EU-27 level in all indicators of the ICRG, as well as in the overall firms
political risk and business regulations. As for the firms freedom indicator, Portugal
lags behind the EU-27 in about 1/2 a point, and an overall reform at this level is able
to increase FDI by almost 40%.17

Table 7: Reforming institutional factors. Impact on FDI, required reform effort, and
efficiency of reform versus the EU-15.
(3) (5) (6) (7)
(1) (2) (4)
(2) − (1) e(4)(3) − 1 |(3)/(1)| (5)/(6)
impact required
index index efficiency
difference coefficient on FDI reform
Portugal EU-15 of reform
(%) effort
Institutional variables
firms freedom∗∗∗ 6.78 7.85 1.07 0.5861 87.40 15.81 5.53
firms political risk∗∗∗ 8.20 8.71 0.51 0.5297 30.96 6.21 4.99
doing business∗∗∗ 7.84 7.92 0.08 0.7127 5.66 0.98 5.75
IEF variables
financial freedom∗∗∗ 5.00 7.29 2.29 0.2439 74.63 45.71 1.63
corruption freedom∗∗∗ 6.50 7.82 1.32 0.2752 43.86 20.33 2.16
business freedom∗∗∗ 7.86 8.63 0.77 0.3128 27.18 9.78 2.78
labor freedom∗∗∗ 4.15 6.45 2.30 0.1514 41.73 55.51 0.75
property rights∗∗∗ 7.00 8.07 1.07 0.1409 16.30 15.31 1.06
investment freedom∗∗∗ 7.00 7.71 0.71 0.1488 11.21 10.20 1.10
trade freedom∗∗∗ 8.66 8.52 -0.14 -0.4289
monetary freedom 8.04 8.29 0.25 0.2306 5.99 3.14 1.91
ICRG variables
corruption∗∗∗ 6.67 7.19 0.52 0.1635 8.89 7.81 1.14
law and order∗∗∗ 8.33 9.05 0.71 0.2567 20.12 8.57 2.35
socioec. conditions∗∗∗ 6.70 7.33 0.63 0.2603 17.82 9.40 1.90
investment profile∗∗ 10.00 9.85 -0.15 0.2457
democratic account.∗∗ 10.00 9.85 -0.15 0.2738
military in politics 10.00 9.64 -0.36 -0.0570
bureaucracy quality 7.50 9.20 1.70 -0.0010
doing business variables
starting a business∗∗∗ 9.12 9.28 0.16 0.7381 12.47 1.75 7.14
protecting investors∗∗∗ 6.08 5.62 -0.46 0.2119
getting credit∗∗∗ 5.56 4.97 -0.59 0.2320
construction permits∗∗∗ 8.34 9.21 0.87 0.4254 44.48 10.37 4.29
closing a business∗∗∗ 8.45 8.45 -0.01 0.1854
trading across borders∗∗∗ 8.84 8.90 0.06 0.4587 2.95 0.72 4.11
enforcing contracts∗∗∗ 7.49 7.88 0.39 0.1856 7.57 5.25 1.44
registering property 7.70 7.96 0.26 0.0671 1.73 3.33 0.52
paying taxes 9.01 9.02 0.01 0.0962 0.08 0.09 0.87
∗ , ∗∗ and ∗∗∗ represent the variables which are significant at 10, 5 and 1 percent significance levels, respectively. The

reform measures are only computed for those variables in which Portugal has an inferior performance relative to the
European Union. The index for the EU-15 does not consider Portugal.

17
However, since the overall freedom score is a simple average of all other factor scores, its efficiency
cannot be directly compared with that of the other indicators.

24
 Table 7 yields slightly different conclusions, as several Portuguese indicators are
more distant from the EU-15 level than from the EU-27 level. Figure provides a
graphical perspective of the effects of reforming Portuguese institutions to the EU-15
level. For the IEF indicators, reforming financial freedom still displays the highest
impact on inward FDI; however, improving corruption freedom is more efficient, and
is able to increase inward FDI in more than 40%. Reforming business freedom leads
to a smaller gain in inward FDI—around 27%—however, this benefit comes at a low
cost, for which it is also a promising area for reform. Reforming other areas—labor
freedom, property rights and investment freedom—can only be done at a very high
cost per unit of impact on FDI. The results for the ICRG variables also suggest
that corruption could be improved; however, the effect of this indicator in inward
FDI is much lower as compared to that of the IEF. Bringing the law and order
indicator to the EU-15 average, on the other hand, is able to increase FDI around
20%, and the reform effort is not significantly high. By the same token, socioeconomic
conditions could also be considered for reform. From the doing business indicators,
Portugal has a better performance than the EU-15 in protecting investors, getting
credit and closing a business, and hence reforms should be discarded in those areas.
Of the remaining areas, reforming the regulatory and administrative burden required
to obtain a construction permit has the highest impact on FDI, and also displays an
efficiency index above 1, being a good candidate for reform. Improving the necessary
bureaucratic steps to start a business is also highly efficient, although the effect is not
as high as for other institutional indicators. The reduced impact on FDI of reforming
other areas of doing business suggest that they should not not be taken as top priorities
for reform.

7 Concluding remarks and policy implications

The purpose of this article was twofold: to analyze the effects of several economic
and institutional factors on bilateral inward FDI for EU countries, and to investigate
which determinants, after reform, are most likely to foster inward FDI into Portugal.
With this intent, to the traditional determinants of FDI we added three additional
sets of institutional variables that: the Index of Economic Freedom, the political risk
index for the International Country Risk Guide, and the Doing Business database.
Our conclusions indicate that, in general, countries with better economic perfor-
mances, better policies, and better institutions, are able to attract more FDI. From
the economic determinants, market size and per capita GDP display the largest im-
pacts on inward FDI, although the effect of the latter is more related with wages

25
 financial freedom (IEF) trading across borders (BD)

construction permits (DB) starting a Business (DB)

corruption freedom (IEF) enforcing contracts (DB)

labor freedom (IEF) corruption (ICRG)

business freedom (IEF) law and order (ICRG)

law and order (ICRG) socioec. conditions (ICRG)

socioec. conditions (ICRG) business freedom (IEF)

property rights (IEF) investment freedom (IEF)

starting a Business (DB) construction permits (DB)

investment freedom (IEF) property rights (IEF)

corruption(ICRG) corruption freedom (IEF)

enforcing contracts (DB) financial freedom (IEF)

trading across borders (BD) labor freedom (IEF)

0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60
impact on FDI (%) reform effort (%)

(a) Impact of reform (b) Required effort for reform

starting a Business (DB)

construction permits (DB)

trading across borders (BD)

business freedom (IEF)

law and order (ICRG)

corruption freedom (IEF)

socioec. conditions (ICRG)

financial freedom (IEF)

enforcing contracts (DB)

corruption (ICRG)

investment freedom (IEF)

property rights (IEF)

labor freedom (IEF)

0 1 2 3 4 5 6 7 8
efficiency of reform (% change in FDI per effort)

(c) Efficiency of reform

Figure 1: Measuring institutional reform in Portugal.

rather than with purchasing power or with the level of development. We also con-
clude that education and corporate taxes play a secondary role in inward FDI relative
to institutions. Within the institutional factors, the independence of the financial
system, the level of corruption, the flexibility of the labor market, the strength and
impartiality of the legal system and the popular observance of the law, and several
business regulations—mainly those related with starting a business, invertor’s protec-
tion, getting credit and licensing procedures for some activities—are the ones which
contribute the most to explain inward FDI. However, from these factors, business
regulations display the highest impact on inward FDI.
However, not all these variables are suited for reform in Portugal. Firstly, be-
cause Portugal may already have good performances in some of these indicators, and
secondly because the marginal effects abstract from the cost of reform. Hence, we

26
complement our analysis by presenting and analyzing 3 measures of institutional re-
form: the impact of the reform on FDI, the required reform effort, and the efficiency of
reform. The results suggest that the most important reform areas to put Portuguese
institutions at the EU-27 level are related with the independence of the financial sys-
tem, and with licensing procedures for some activities. A more ambitious reform plan
which puts Portuguese institutions at the EU-15 average level should also consider
reforms aimed at decreasing corruption, at improving the strength and impartiality of
the legal system and the popular observance of the law, at developing socioeconomic
conditions, and at simplify the necessary procedures to start a business. Increasing
labor market flexibility also has a large impact on inward FDI, but the required reform
effort makes the reform unattractive.

A Data description
Distance and economic variables
The following variables are used.
FDI stock: Bilateral inward FDI stocks in Millions of Euros. Source: Eurostat.
FDI flows: Bilateral FDI inflows in Millions of Euros. Source: Eurostat.
Border: Dummy variable which takes the value of 1 if the source and host country
share a common border and 0 otherwise.
Distance: Distance, in Kilometers, from the capital of the source country to the
capital of the host country, calculated using the great circle distance.
GDP: Gross Domestic Product in Millions of Euros at current market prices. Source:
Eurostat.
GDP per capita: Per capita GDP in Thousands of Euros at current market prices.
Source: Eurostat.
GDP growth: Real GDP growth rate (percentage) relative to previous year. Source:
Eurostat.
Openness: Degree of openness, measured by the ratio of exports plus imports over
GDP. Source: Eurostat.
Telephone lines: Number of telephone lines per 100 people. Source: World Bank.
Secondary education: Percentage of population aged 25 to 64 having completed
tertiary education. Source: Eurostat.
EATR: Effective average tax rate in the host country (percentage). The data was

27
kindly provided by Michael Overesch (see Overesch and Rincke, 2009).

Index of Economic Freedom

The Index of Economic Freedom is computed by the Heritage Foundation, and the
data can be found in their website (www.heritage.org/index). The overall index of
economic freedom is constructed by taking the simple average across 10 different
indexes. The indexes, described below, were rescaled to the 0–10 range to ease the
exposition. A higher score means a better performance.
Business Freedom: Measures the ability to start, operate, and close a business,
thus representing the overall burden from regulation and government efficiency in the
regulatory process. This score is constructed using data on 10 factores from the Doing
Business study: number of procedures to start a business, number of days to start
a business, cost of starting a business, minimum capital required to start a business,
number of procedures to obtain a licence, time in obtaining a licence, cost in obtain-
ing a licence, time to close a business, cost of closing a business, and recovery rate in
closing a business.
Trade Freedom: Composite measure of the absence of tariff and non-tariff barriers
that affect imports and exports of goods and services. It is based on the trade-weighted
average tariff rate and on non-tariff barriers.
Fiscal Freedom: Measures the tax burden imposed by the government. It includes
the direct tax burden on individual and corporate incomes, and tax revenues as a
percentage of GDP.
Government Freedom: Accesses the weight of government expenditures on GDP.
According to this index, a higher share of government expenditures on GDP is asso-
ciated to a lower score.
Monetary Freedom: Measures price stability (the weighted average inflation rate
for the last 3 years) and price controls. The index is constructed such that inflation
and price controls originate lower scores.
Investment Freedom: Evaluates a variety of restrictions imposed on investment,
namely restrictions on FDI, transparency and bureaucracy in the foreign investment
code, restrictions on land ownership, sectoral investment restrictions, expropriation
of investments without fair compensation, foreign exchange controls and capital con-
trols.
Financial Freedom: Measures bank security and the independence of the financial
system from government control. This score comprises the extent of government regu-

28
lation in the financial system, the extent of state intervention in banks, the difficulty of
operating financial services, and the government influence on the allocation of credit.
Property Freedom: Accesses the ability of individuals to accumulate private prop-
erty, secured by clear laws that are fully enforced by the state. Thus, it measures the
extent to which laws protect private property, the degree of enforcement of those laws
by the government, and the likelihood of expropriation. It also analyzes the indepen-
dence of the judiciary system, the level of corruption within the judiciary system, and
the ability of individuals and business men to enforce contracts.
Freedom from Corruption: Measures the degree of corruption within a country.
Labor Freedom: Measures several aspects of the legal and regulatory framework of
a country’s labor market, such as regulations on the minimum wages, laws inhibiting
layoffs, regulatory burdens on hiring, among others.

International Country Risk Guide—Political Risk Rating

The International Country Risk Guide computes 22 variables in three categories of
risk: political, financial, and economic. The political risk rating includes 12 variables,
covering political and social attributes. Although some of them are more related
with inward FDI than others, all are used in the principal components analysis, as
explained in the main text. In order to facilitate the comparison in the magnitude of
the coefficients, we convert all indexes to the 0–10 range.
Government stability: Accesses the government’s ability to carry out its declared
program and the ability to stay in office. It is constructed from 3 subcomponents:
government unity, legislative strength and popular support.
Socioeconomic conditions: Accesses the pressures that could constraint govern-
ment action or fuel social dissatisfaction. It is constructed from 3 subcomponents:
unemployment, consumer confidence and poverty.
Investment profile: Accesses the risks to investment that are not covered by other
economic, political and financial risk components. It is constructed from 3 subcom-
ponents: contract viability/expropriation, profile repatriation and payment delays.
Internal conflict: Accesses the political violence and its potential impact on gov-
ernance. It is constructed from 3 subcomponents: civil way/coup threat, terror-
ism/political violence and civil disorder.
External conflict: Accesses the risk to the government from foreign action, both
originating from non-violent pressures (for instance diplomatic pressures, trade re-
strictions, territorial disputes) and violent pressures (cross-border conflicts and war).

29
It is constructed from 3 subcomponents: war, cross-border conflict and foreign pres-
sures.
Corruption: Accesses corruption in the political system, including bribes, exchange
controls and tax assessments, among others.
Military in politics: Measures the involvement of military in politics, which distorts
government policy and diminishes democratic accountability.
Religious tensions: Measures the involvement of religious groups in politics. Re-
ligious groups often seek to replace civil laws by religious laws, thus distorting and
constraining government action.
Law and order: Measures the strength and impartiality of the legal system and the
popular observance of the law.
Ethnic tensions: Measures racial, nationality and language tensions, which origi-
nate intolerance and unwillingness to make compromises.
Democratic accountability: Measures the responsiveness of government to its peo-
ple. The score for this component is based on the following types of governance: alter-
nating democracy, dominating democracy, de-facto one-party state, de jure one-party
state and autarchy.
Bureaucracy quality: Measures the institutional strength and quality of the bu-
reaucracy, and the extent to which bureaucracy is autonomous from the political
pressure and has a established mechanism for recruitment and training.

Ease of Doing Business

The World Bank’s Doing Business (DB) database measures business regulations and
therefore the cost of a firm operating in a country. More specifically, it provides
quantitative assessment for starting a business, dealing with construction permits,
registering property, getting credit, protecting investors, paying taxes, trading across
borders, enforcing contracts and closing a business, in a total of 33 raw factores,
described below. For the analysis, these factors were aggregated into 9 indexes ranging
from 0 to 10, each summarizing a specific topic of doing business. To compute these
indexes, we proceeded as follows. First, we represented each raw factor by an index,
ranging from 0 to 10, according to the min-max standardization method,

factork − factormin
Scorek = 10
factormax − factormin

30
if higher factor values imply better performances (e.g., strength of legal rights, recov-
ery rate when closing a business), or

factork − factormin
Scorek = 10 − 10
factormax − factormin
if higher factor values imply worst performances (e.g., procedures, time, cost).
This was done for all countries in the DB database. According to these formulas,
all scores are organized such that higher values always mean better performances. The
topic score is the simple average of all factors that compose that topic. An overall
ease of doing business index is created by taking the simple average of the 9 topic
scores.

Starting a business
The starting a business index measures all procedures, costs and time that are for-
mally required for an entrepreneur to start up and formally operate an industrial or
commercial business. It includes the following factors:

Procedures: Number of interactions of the company founders with external parties

in order to formally start operating a business.
Time: Median duration in calendar days that is necessary to complete a procedure
with minimum follow-up with government agencies and no extra payments.
Cost: All official fees and fees for legal or professional services if such services are
required by law in order to start operating a business. It is recorded as a percentage
of the economy’s income per capita.
Minimum Capital: The amount that the entrepreneur needs to deposit in a bank
or with a notary before registration and up to 3 months following incorporation in
order to start operating a business. It is recorded as a percentage of the economy’s
income per capita.

Dealing with construction permits

The dealing with construction permits index measures all procedures required for a
business in the construction industry to build a standardized warehouse, as well as the
costs and time required to complete the procedures. It includes the following factors:

Procedures: Number of procedures required for a business in the construction in-

31
dustry to build a standardized warehouse.
Time: Median duration in calendar days that is necessary to complete the required
procedures.
Cost: All fees associated with completing the procedures to legally build a warehouse.
The cost is recorded as a percentage of the economy’s income per capita.

Registering property
The registering a property index records the necessary procedures that a business
man must incur to purchase a property from another business man and to transfer
the property title to his name, as well as the associated costs and time. It includes
the following factors:

Procedures: Number of procedures that are legally or in practice required for regis-
tering a property.
Time: Median duration that property lawyers, notaries or registry officials indicate
is necessary to complete the procedures for registering a property.
Cost: All the necessary fees to register a property. This variable is recorded as a
percentage of the property value.

Getting credit
The getting credit index measures the legal rights of borrowers and lenders with re-
spect to secured transactions and the sharing of credit information. It includes the
following factors:

Strength of legal rights: Index that measures the degree to which collateral and
bankruptcy laws protect the rights of borrowers and lenders and thus facilitate lending.
It ranges from 0 to 10.
Depth of credit information: Index that accesses the rules and practices affecting
the coverage, scope and accessibility of credit information, regardless of whether this
information is available through a public credit registry or through a private credit
bureau. It ranges from 0 to 6.
Public registry coverage: Reports the number of individuals and firms listed in
a public credit registry with information on their borrowing history from the past 5
years. It is measured as a percentage of adults aged 15 and above.

32
Private bureau coverage: Reports the number of individuals and firms listed by
a private credit bureau with information on their borrowing history from the past 5
years. The number is expressed as a percentage of the adult population aged 15 and
above.

Strength of investor protection

The strength of investor protection index measures the strength of minority share-
holder protections against directors misuse of corporate assets for personal gain. It
includes the following factors:

Extent of disclosure: Index which accesses who can approve related-party transac-
tions and the requirements for external and internal disclosure in case of related-party
transactions. It ranges from 0 to 10.
Extent of director liability: Index which measures the ability of shareholders to
hold the interested party and the approving body liable in case of a prejudicial related-
party transaction, the availability of legal remedies (damages, repayment of profits,
fines, imprisonment and rescission of the transaction) and the ability of shareholders
to sue. It ranges from 0 to 10.
Ease of shareholder suits: Index that measures the documents and information
available during trial and the access to internal corporate documents. It ranges from
0 to 10.

Paying taxes
The paying taxes index measures the tax burden and mandatory contributions that a
medium size company must pay in a given year, as well as the administrative burden
of paying taxes and contributions. It includes the following factors:

Payments: Reflects the total number of taxes and contributions paid, including con-
sumption taxes, as well as the method of payment, the frequency of payment, the
frequency of filing, for a company during the second year of operation.
Time: Measures the hours per year a company spends to prepare, file and pay the
corporate tax, the value added tax, and social contributions. It includes the time
spent to collect information and to compute the amount payable.
Total tax rate: Measures all taxes and contributions (corporate taxes, social contri-
butions, labor taxes, property taxes, and other taxes) paid by firms as a percentage

33
of total profits.

Trading across borders

The trading across borders index measures procedural requirements for exporting and
importing a standardized cargo of goods by ocean transport. It includes the following
factors:

Documents to export: Number of bank documents, customs clearance documents,

port and terminal handling documents and transport documents required for export-
ing.
Documents to import: Number of bank documents, customs clearance documents,
port and terminal handling documents and transport documents required for import-
ing.
Time to export: Time (in calendar days) to obtain all documents required for in-
land transport and handling, for customs clearance and inspections, and for port and
terminal handling, for exporting a standardized cargo.
Time to import: Time (in days) to obtain all documents required for to inland
transport and handling, for customs clearance and inspections, and for port and ter-
minal handling, for importing a standardized cargo.
Cost to export: Measures the costs (in US dollars per container) of all documenta-
tion, inland transport and handling, customs clearance and inspections and port and
terminal handling, for exporting a standardized cargo.
Cost to import: Measures the costs (in US dollars per container) of all documenta-
tion, inland transport and handling, customs clearance and inspections, and port and
terminal handling, for importing a standardized cargo.

Enforcing contracts
The enforcing contracts index measures the efficiency of the judicial system in resolv-
ing a commercial dispute. It includes the following factors:

Procedures: Number of procedures resulting from a comercial dispute, either be-

tween the parties or between them and the court officer. It comprises the steps to file
the case, the steps for trial and judgment and the steps to enforce the judgment.
Time: Number of calendar days, from the moment the lawsuit is filed on court until

34
payment. It includes the necessary time to file and serve the case, the time od the
trial, and the time to enforce the judgment.
Cost: Average attorney fees, court costs (including expert fees) and enforcement costs
a firm must incur if a commercial dispute goes to trial. It is measured as a percentage
of claims.

Closing a business
The closing a business index measures the time, cost and outcome of insolvency pro-
ceedings involving domestic entities. It includes the following factors:

Time: Calendar years required for creditors to recover their credit.

Cost: Court fees, fees of insolvency administrators, lawyers’ fees, and assessors’ and
auctioneers’ fees required to close a business. It is measure as a percentage of the
debtor’s estate value.
Recovery rate: Measures the percentage recovered by creditors, i.e., the present
value of debt than can be recovered.

References
Altomonte, C. and C. Guagliano, “Comparative study of FDI in Central and Eastern
Europe and the Mediterranean,” Economic Systems 27 (2003), 223–246.
Bajo-Rubio, O. and S. Sosvilla-Rivero, “An Econometric Analysis of Foreign Direct
Investment in Spain, 1964–89,” Southern Economic Journal 61 (1994), 104–120.
Barrell, R. and N. Pain, “An Econometric Analysis of U.S. Foreign Direct Investment,”
The Review of Economics and Statistics 78 (1996), 200–207.
———, “Foreign direct investment, technological change, and economic growth within Eu-
rope,” The Economic Journal 107 (1997), 1770–1786.
Bellak, C., M. Leibrecht and J. Damijan, “Infrastructure endowment and corpo-
rate income taxes as determinants of Foreign Direct Investment in Central and Eastern
European Countries,” LICOS Discussion Paper (2007).
Bénassy-Quéré, A., M. Coupet and T. Mayer, “Institutional determinants of foreign
direct investment,” The World Economy 30 (2007), 764–782.
Bevan, A. and S. Estrin, “The Determinants of Foreign Direct Investment into European
Transition Economies,” Journal of Comparative Economics 32 (2004), 775–787.

35
Bhattacharya, R., I. Patnaik and A. Shah, “Export versus FDI in services,” IMF
Working Paper WP/10/290 (2010).
Billington, N., “The Location of Foreign Direct Investment: An Empirical Analysis,”
Applied Economics 31 (1999), 65–76.
Biswas, R., “Determinants of foreign direct investment,” Review of Development Eco-
nomics 6 (2002), 492–504.
Blonigen, B., “Firm-Specific Assets and the Link between Exchange Rates and Foreign
Direct Investment,” The American Economic Review 87 (1997), 447–465.
Borensztein, E., J. De Gregorio and J. Lee, “How does foreign direct investment
affect economic growth?,” Journal of international Economics 45 (1998), 115–135.
Campa, J. M., “Entry by Foreign Firms in the U.S. Under Exchange Rate Uncertainty,”
Review of Economics and Statistics 75 (1993), 614–622.
Cassou, S., “The Link Between Tax Rates and Foreign Direct Investment,” Applied Eco-
nomics 29 (1997), 1295–1301.
Chakrabarti, A., “The determinants of foreign direct investment: sensitivity analysis of
cross-country regressions,” KYKLOS 54 (2001), 89–114.
Culem, C., “The Locational Determinant of Direct Foreign Investment Among Industrial-
ized Countries,” European Economic Review 32 (1988), 885–904.
Cushman, D., “Real Exchange Rate Risk, Expectations, and the Level of Direct Invest-
ment,” Review of Economics and Statistics 67 (1985), 297–308.
de Mooij, R. and S. Ederveen, “Taxation and Foreign Direct Investment: A Synthesis
of Empirical Research,” International Tax and Public Finance 10 (2003), 673–693.
Devereux, M. and R. Griffith, “Taxes and the Location of Production: Evidence from
a Panel of US Multinationals,” Journal of Public Economics 68 (1998), 335–367.
Dunning, J., Multinational enterprises and the global economy (Cheltenham: Edward El-
gar, 2008).
Eaton, J. and A. Tamura, “Bilateralism and regionalism in Japanese and US trade and
direct foreign investment patterns,” NBER Working Paper 4758 (1995).
Flamm, K., “The Volatility of Offshore Investment,” Journal of Development Economics
16 (1984), 231–248.
Froot, K. and J. Stein, “Exchange Rates and Foreign Direct Investment: An Imperfect
Capital Markets Approach,” The Quarterly Journal of Economics 106 (1991), 1191–1217.
Gao, T., “Foreign direct investment from developing Asis: Some distinctive features,”
Economics Letters 86 (2005), 29–35.
Goldberg, L. and C. Kolstad, “Foreign Direct Investment, Exchange Rate Variability
and Demand Uncertainty,” International Economic Review 36 (1995), 855–873.

36
Grubert, H. and J. Mutti, “Taxes, Tariffs and Transfer Pricing in Multinational Cor-
porate Decision Making,” The Review of Economics and Statistics 73 (1991), 285–293.
Hartman, D., “Tax Policy and Foreign Direct Investment in the United States,” National
Tax Journal 37 (1984), 475–487.
———, “Tax Policy and Foreign Direct Investment,” Journal of Public Economics 26 (1985),
107–121.
Head, K., J. Ries and D. Swenson, “Agglomeration Benefits and Location Choice:
Evidence from Japanese Manufacturing Investments in the United States,” Journal of
International Economics 38 (1995), 223–247.
Helpman, E., M. J. Melitz and S. R. Yeaple, “Export versus FDI with heterogeneous
firms,” The American Economic Review 94 (2004), 300–316.
Herzer, D., “The long-run relationship between outward FDI and domestic output: Evi-
dence from panel data,” Economics Letters 100 (2008), 146–149.
———, “Outward FDI and economic growth,” Journal of Economic Studies 37 (2010),
476–494.
Hines, J., “Altered States: Taxes and the Location of Foreign Direct Investment in Amer-
ica,” The American Economic Review 86 (1996), 1076–1094.
Hines, J. and E. Rice, “Fiscal Paradise: Foreign Tax Havens and American Business,”
The Quarterly Journal of Economics 109 (1994), 149–182.
Hwang, J. G. and D. Nettleton, “Principal components regression with data chosen
components and related methods,” Technometrics 45 (2003), 70–79.
Janicki, H. and P. Wunnava, “Determinants of Foreign Direct Investment: Empirical
Evidence From EU Accession Candidates,” Applied Economics 36 (2004), 505–509.
Klein, M. and E. Rosengren, “The Real Exchange Rate and Foreign Direct Investment
in the United States: Relative Wealth vs. Relative Wage Effects,” Journal of International
Economics 36 (1994), 373–389.
Knack, S. and P. Keefer, “Institutions and Economic Performance: Cross-country Tests
Using Alternative Institutional Measures,” Economics and Politics 7 (1995), 207–227.
Kogut, B. and S. J. Chang, “Platform Investments and Volatile Exchange Rates: Direct
Investment in the U.S. by Japanese Electronic Companies,” Review of Economics and
Statistics 78 (1996), 221–231.
Kravis, I. and R. Lipsey, “The Location of Overseas Production and Production for
Exports by US Multinational Firms,” Journal of International Economics 12 (1982),
201–223.
Lee, J.-Y. and E. Mansfield, “Intellectual Property Protection and U.S. Foreign Direct
Investment,” The Review of Economics and Statistics 78 (1996), 181–186.

37
Loree, D. W. and S. E. Guisinger, “Policy and non-policy determinants of U.S. equity
foreign direct investment,” Journal of International Business Studies 26 (1995), 281–299.
Love, J. H. and F. Lage-Hidalgo, “Analysing the determinants of US direct investment
in Mexico,” Applied Economics 32 (2000), 1259–1267.
Overesch, M. and J. Rincke, “The Dynamics of Corporate Tax Setting in Europe,
1984–2006,” CESifo Working Paper No. 2535 (2009).
Scaperlanda, A. and R. Balough, “Determinants of U.S. Direct Investment in the
E.E.C.: Revisited,” European Economic Review 21 (1983), 381–390.
Schmitz, A. and J. Bieri, “ECC Tariffs and U.S. Direct Investment,” European Economic
Review 3 (1972), 259–270.
Schneider, F. and B. Frey, “Economic and Political Developments of Foreign Direct
Investment,” World Development 13 (1985), 161–175.
Stein, E. and C. Daude, “Longitude matters: Time zones and the location of foreign
direct investment,” Journal of International Economics 71 (2007), 96–112.
Stevens, G., “Politics, Economics and Investment: Explaining Plant and Equipment
Spending by U.S. Direct Investors in Argentina, Brazil and Mexico,” Journal of Interna-
tional Money and Finance 19 (2000), 153–183.
Swenson, D., “The Impact of U.S. Tax Reform on Foreign Direct Investment in the United
States,” Journal of Public Economics 54 (1994), 243–266.
Tavares, J., “Institutions and economic growth in Portugal: a quantitative exploration,”
Portuguese Economic Journal 3 (2004), 49–79.
Tsai, P.-L., “Determinants of Foreign Direct Investment and its Impact on Economic
Growth,” Journal of Economic Development 19 (1994), 137–163.
Walsh, J. and J. Yu, “Determinants of foreign direct investment: A sectoral and institu-
tional approach,” IMF Working Paper WP/10/187 (2010).
Wei, S.-J., “How Taxing is Corruption on International Investors,” Review of Economics
and Statistics 82 (2000a), 1–11.
———, “Local Corruption and Global Capital Flows,” Brookings Papers on Economic Ac-
tivity 2 (2000b), 303–346.
Wheeler, D. and A. Mody, “International Investment Location Decisions: The Case of
U.S. Firms,” Journal of International Economics 33 (1992), 57–76.

38