Foreign direct investment in non-landlocked Africa countries

Abstract

This paper undertook a comparative static analysis of
foreign direct investment (FDI) determinants on FDI inflows to
Arabophones and Francophones countries in Africa. This paper used
an unbalanced panel data on non-landlocked Africa member states
excluding the Lusophone member states from 1980 to 2010. To
accomplish this, Panel unit root test, Pedroni (Engle-Granger
Based) Panel Co-integration Test, multi-collearity test and
Hausman test were conducted before using a panel dynamic ordinary
least square estimation under the fixed effect specification for
the entire models. Contemporaneous trade openness, financial
development, economic growth and short- term investment in
infrastructural was positive determinants of FDI inflows to
non-landlocked Africa countries excluding the Lusophone Africa.
On the contrary, lagged of infrastructure development, lead of
domestic savings and balance of payment was identified as part of
FDI determinants to Arabophones member states, even though the
effects were negative. Nevertheless, the result showed that
language is not a significant factor for attracting FDI inflows.
Finally, policies/strategies to improve the economic growth and
trade openness should be put in place in order to attract more
FDI inflows to non-landlocked Africa.

Key words: Foreign direct investment,
non-landlocked, Arobophones and Francophones.

Introduction

Developing countries are ever more informed on the
significant role of foreign direct investment
(FDI[1]as an engine of growth in their economies.
Theories have elicited that FDI can induce economic growth by
providing essential capital and skills, participating in large
projects and as well as being a vehicle for technology transfer.
For many developing countries, FDI has been an essential
mechanism that promotes industries, entailed the spirit of
competition and in the long-run makes them have a potential
comparative advantage over other economies (Addison et
al, 2004; UNCTAD, 2003; Dunning & Hamdani,
1997).

Literatures have identified many determinants of FDI in
an economy. These determinants effects differ from one economy or
region to another. According to Calvo et al. (1993)
study argued that, the total FDI inflows in a country are stirred
largely by push factors, such as economic growth and return on
investment in industrial countries. That is, the foreign
investors" main objective is profit maximization. The implication
is that investors will move to a new destination if the return on
investment in that destination is promising. Apart from the FDI
determinants associated with push factors, there are other
determinants of FDI which are linked to pull factors. According
to Collins (2002), the Pull factors enabled the investors to
choose a profitable investment allocation among developing
countries.

In addition to pull and push factors, foreign direct
investor"s reason of investing in a particular economy might be
related to their motives. For example, a natural-resource-seeking
investor aims to exploit the natural resource endowments of that
country. This investor will mostly channel funds to companies
extracting oil (in Nigeria, and Cote d"Ivoire), gold (in Ghana)
and diamond (in Botswana), etc. Other types of investor are:
Market-seeking investors who aim at taking advantage of new
markets in terms of their sizes and/or growths.
Efficiency-seeking investors take advantage of special features
such as the costs of labor, the skills of the labor force, and
the quality and efficiency of infrastructure. Lastly,
strategic-asset-seeking FDI investors situate at a place where
they can take advantage of what is readily available in terms of
research and development and other benefits.

Based on the empirical and theoretical identification of
various FDI determinants work covering a wider range of countries
and its effect on FDI to recipient economy, which might depend
not only on local conditions and policies but also the linguistic
approach cognition. Simply put, one question this paper
postulated was whether the disparities in terms of language were
a factor in attracting FDI inflows to Northern Africa. That is,
whether the official language was a latent determinant
instrumented for the economy in question to harvest benefits
derived from FDI inflows.

In this paper, the author explores a comparative static
analysis of FDI determinants effect on FDI inflows to Arabophones
and Francophones countries in Africa. In order to do this, the
author sourced data from UNCTAD"s and Africa Development
indicator database. In total, the sample covered 12 African
countries, among which five of were Arabophones countries,
namely, Algeria, Egypt, Morocco, Tunisia and Mauritania while the
rest, Gabon, Benin, Guinea, Madagascar, Senegal and Côte
d'Ivoire belonged to Francophones countries. The paper adopted a
dynamic ordinary least square (DOLS) method to achieve its entire
objectives. The rest of the paper was organized as follows,
Section two contained a comparative static analysis overview
performances of FDI determinants in Arabophones and Francophone
economy, while section three reviewed the theoretical and
empirical literature from other related studies, from which the
paper derive determinants of FDI having a potential impact on FDI
inflows Arabophones and Francophones countries. The methodology,
econometric specification and estimation strategy were presented
in section four. Empirical results and their interpretations were
discussed in section five, while section six summarized, and
present policy discussion and made suggestions for further
research.

Overview
performances of fdi determinants in arabophones and francophone
economy

Table1: Comparative Static Analysis

Sources: UNCTAD and WDI and the
calculation were done by the author.

In table 1, both the Arabophones and Francophones member
states on average were running balance of payment deficits from
over the period of this empirical study. For example, between
1980 and 1990, the BOP deficits for Arabophones and Francophones
were 6.02 percent and 6.51 percent of their gross domestic
product respectively. But the deficit on average reduced to 0.88
percent between 1991 and 2000 and further to 0.74 percent from
2001 and 2010 for Arabophones member states economy. There were
reductions on average in Francophone countries within the same
periods but not as compared with Arabophones
countries.

In both Arabophones and Francophone member states, their
domestic savings increase but Arabophones domestic savings
increases more than Francophone over the same periods. The
financial development (FD) in Arabophones, were more than twice
liberalized than in Francophone member states. For instance, it
was 51.20 percent and 20.72 percent for Arabophones and
Francophones in 1980-1990 respectively. The development of the
financial system dropped a bit for both member states in
1991-2000 and picked up to 56.19 percent and 23.01 percent in
2001-2010 for Arabophones and Francophones
respectively.

Furthermore, there is fairly level of trade openness for
both Arabophones and Francophones member states, but Francophones
countries were more open than Arabophones countries in 1980-1990
and 1991-2010. This is an evidence of adopting globalization in
their various economies. The government participation in both
member states were low and fairly stable in both economies.
Domestic investments in Arabophones countries were slightly
higher when compared to Francophones countries. Finally, the FDI
inflows in both countries were approximately the same over three
deciles observation.

In figure 1 below, FDI inflows in Arabophones were
increased from 1980 to 1994, and dropped from 1994-1995 while
over the same period FDI inflows to Francophones countries were
fairly stable. The FDI inflows for both member states were
increased from 2004 to 2006, and Arabophones attracts more than
Francophones member states. From 2006 to 2008, FDI inflows to
Arabophones countries were approximately stable but the increase
of FDI inflows continued flowing to Francophones countries over
the same period. Finally, after 2008, the FDI inflows to
Arabophones declined.

Figure 1: FDI inflows in Arabophones and
Francophone"s countries in Africa (1980-2010)

Source: Africa development Indicators
various years

Literature
review

3.1 Reviews of Theoretical Literature

There are a number of theories that have explained
Foreign Direct Investment inflows (FDI) in Africa. FDI theories
are mainly based on imperfect market conditions and imperfect
capital market. Some of these theories consider the effect of
non-economic factors on FDI while others explain the emergence of
Multinational Corporations (MNCs) exclusively among developing
countries. For instance, neo-classical economic theory assumes a
free capital markets and diminishing returns. According to the
theory, capital should flow from capital abundant countries
(developed countries) to capital scarce countries (developing
countries) until the marginal returns to capital in both
countries are equal. Such flows can contribute immensely to
closing domestic savings gap in developing countries (Page &
Velde, 2004). Indeed, Mundell"s (1957) in the Heckscher-Ohlin
theory of trade postulated that capital will move from rich
countries to poor countries. The implication is that, it is more
profitable to invest capital where it is scarce than where it is
abundant. It holds on the assumption that there should be no
outflows of FDI from developing countries.

In addition, FDI is associated to production as well as
capital flows, which are influenced by other factors. From a
conventional trade point of view, trade and FDI might be viewed
as substitutes in absence of the influence of factors such as
technology and firm-specific assets; otherwise they may be seen
as complements (Markusen, 1984 & 1995). A notable example of
firm-specific assets is brand names which are acquired via
advertising or firm specific knowledge acquired via Research and
Development (R&D).

Nevertheless, there are other factors that influence FDI
inflows in a particular country other than differences in factor
endowments and factor prices. Trade economists believe that
return on investment, imperfect competition and product
differentiation and other factors related to the comparative
advantage paradigm are part of FDI investor"s motive. Apart from
trade theory, Dunning"s eclectic paradigm explained in detail
some important variables essential in attracting FDI inflows to
Africa.

Dunning"s eclectic paradigm which is the combination of
the imperfect market-based theories of FDI, that is, industrial
organization theory, internalization theory and location theory.
It postulates that, at any given time, the stock of foreign
assets owned by a multinational firm is determined by a
combination of firm specific or ownership advantage, the extent
of location bound endowments, and the extent to which these
advantages are marketed within the various units of the firm. The
theory further argued that, the reason why FDI inflows are
greater in one country but not in another is the country"s
locational advantage. (Dunning, 1980, 1993). Dunning identifies
four major types of FDI investors and their associated motives
which are; market-seeking foreign investors will move to a
country with less openness, efficiency- seeking foreign investors
will move to a country with low labor cost, while natural
resource-seeking foreign investors will invest on a country that
is endowed with natural resources and strategic asset seeking
investors flows to a country that is advanced in technology,
skills or take over brand names.

Based on the foreign direct investment theories reviewed
so far, it is improper to accept any of the above theories to
explain the effect of some selected determinants of foreign
direct investment inflows on Arabophones and Francophone"s
countries. This is because the above theories was not able to
combine both the investors" motives related factors and that of
the host countries related factors detrimental to foreign direct
investment inflows. For this reason, the current paper adopted
the integrative theory.

The integrative theory gives a clearer view of the
effects of some determinants of FDI inflows by analyzing it from
the perspectives of host countries as well as investors. It
incorporated the effect of the short-run, contemporaneous and the
long-run effect of the FDI determinants on FDI inflows into its
theory. For instance, eclectic paradigm, which is the combination
of the firm and internalization theories, and industrial
organization theory tackle FDI determinants from the viewpoint of
the firm. The neoclassical and perfect market theories examine
FDI from the perspective of free trade. The development and
dependency theories shed light on the perspective of the host
nation while integrative theory integrates those neoclassical,
developments, dependency and eclectic paradigm that are important
in attracting FDI inflows into a country in its
theory.

According to integrative theory, foreign direct inflows
are a function of the host country factors, the firms" factors
and the foreign direct investors" related factors. Integrative
theories account for the multiplicity of heterogeneous variables
involved in attracting the FDI inflows in Africa. It also forms
the current study theoretical background for identifying
determinants of foreign direct investment using linguistic
approach for comparative static analysis. Integrative theory is
presented as;

FDI inflows = F (factors related to the host
country, firm specific factors and investor factors related to
their motives).

The interactive theoretical framework gives the basis
for adopting dynamic ordinary least squares (DOLS) model of the
current study to identify determinants of foreign direct
investment in Arabophones and Francophone"s countries in Africa.
The model captured all the relevant factors identified by
interactive theory.

3.2 Reviews of Empirical studies

There are a number of empirical studies that have shown
the effect of determinants of FDI in attracting it to the host
countries. FDI has empirically been found to stimulate economic
growth by a number of researchers (Glass & Saggi, 1999). It
was also true according to Dees (1998) that China"s economic
growth is an evidence of FDI spillovers. In Latin America FDI
inflows has contributed a lot to the economic development of
selected Latin America economy (De Mello, 1997). Some other
studies conducted in the same region found that a certain income
level is a crucial determinant of FDI inflows; below that level
it does not have any effect (Blomstrom et al., 1994,
Bengos & Sanchez-Robles, 2003). Their explanation was that a
host country should have the capacity to absorb the effect of FDI
inflows otherwise it would have no effect on the host country"s
economy.

Neoclassical economists have argued that FDI influences
economic growth by increasing the amount of capital per person.
It does not influence long-run economic growth due to diminishing
returns to capital. A study conducted in East Asia has shown that
FDI has a positive effect on less advanced economy"s output than
advanced economy (Bende-Nanende et al. 2002). However,
there was no consensus on the positive impact of FDI on growth.
For instance, a study conducted by Globeram (1979) on some
selected developing countries stipulated that FDI has a
significant effect on economic growth of the countries studied
via its positive effect on domestic firms. Eighteen years later,
Regnanet (1997) conducted a study in the same countries which he
found the same result as Globeram. On the contrary, a study
conducted in Asia countries by Aitken et al. (1997)
found a negative relationship between FDI and economic growth.
Other found inconclusive result. Zhang (2001) emphasizes that the
effect of FDI has on the growth or growth of FDI of any economy
may be country and period specific. Another determinant of FDI is
financial development.

Financial Development stimulates the growth process of
an economy. Its indicators assess the size, activities, and
efficiency of a financial intermediaries and market in a country
(Beck, Demirguc-Kunt & Levine, 2000). Financial development
can make it easier for aid recipient countries to assess aid from
foreign aid donors (Nkusu and Sayek, 2004). Furthermore, Beck"s
study shows that countries with an effective financial sector
have a comparative advantage in manufacturing industries.
Nabamita and Sanjukwa (2008) agreed with Beck and highlighted the
functions of financial development as the following; channeling
resources efficiently, mobilizing savings, reducing the
information asymmetry problem, facilitating trading, hedging,
pooling and diversification of risk, aiding the exchange of goods
and services and monitoring managers by exerting corporate
control. Hermes and Lensink (2003) argued that financial
development has a positive effect on FDI. The reasons advanced
for the result were; a developed financial system mobilizes
savings efficiently and as such may increase the amount of
resources available to undertake investment. A well developed
financial system reduces financial transaction and information
acquisition costs. Financial development also speeds up the
adoption of new technologies by minimizing the risk associated
with it. With a well developed financial system, foreign
investors are able to deduce how much they can borrow for
innovative activities and are able to make investment financing
decision ahead of time. It also increases liquidity and, thus,
facilitates trading of financial instruments and timing and
settlement of such trades (Levine, 1997). This will also lead to
greater FDI inflows as the projects can be undertaken with lesser
time being spent in settling the trades. Nabamita and Sanjukta
(2008) conducted a study on developing countries and their
findings were that there was a positive relationship between
financial development and FDI inflows after a threshold level of
financial development is reached. Beyond that level the effect
becomes negative.

In theory, openness is one of the determinants of FDI
inflows. Its effect on FDI inflows to an economy differs based on
the investor"s motivation for engaging in FDI activities
(Brainard, 1997; Markusen & Maskus, 2002; Navaretti and
Venables, 2004). The more open an economy becomes the better for
non-market seeking investors who would like to use the
destination as an export base. On the contrary, market seeking
investor who"s their target market is the host countries would
prefer less openness. There are vast numbers of empirical studies
that have found openness as an important determinant of FDI
inflows. Chakrabarti (2001) study on the determinants of FDI
inflows to developing countries, the study found out that
openness is an important determinant of FDI and is positively
related to FDI inflows. Moosa and Cardak (2006) conducted a
similar study as Chakrabarti"s, they discovered that export as a
percentage of GDP positively affects FDI inflows. Openness is
found to be positively and significantly related to FDI inflows
in developing countries (Lucas, 1993, Singh & Jun 1995). On
the contrary, Busse & Hefeker (2007) and Globerman and
Shapiro (2002), concluded that openness is statistically
insignificant and does not affect FDI inflows. Some other
researchers found an inclusive result with respect to FDI, such
as Goodspeed et al. (2006). They found a positive and
significant relation with FDI inflows in one model and
insignificant in other specifications of the empirical model.
When testing the vertical, horizontal and knowledge capital
models, Markusen and Maskus (2002) concluded that trade precincts
might be less significant as a motivation for horizontal
tariff-jumping investments in developing countries. This means
that a greater degree of openness will have less of an effect on
the market-seeking investments in developing countries in
comparison to developed countries.

Asiedu (2002) conducted a study by selecting some
African countries; the author"s result implied that openness on
FDI has a lesser effect in Sub-Saharan Africa when compared to
other developing countries. Contrary to Asideu, Tøndel
(2008) found that openness has a higher influence of FDI inflows
in Sub-Saharan Africa than for other countries. There is also
some evidence with respect to differences within the group of
transition countries. When the economies first opened up for
foreign participation in the 1990s, investments in Central Europe
were vertical, whereas FDI activities in the Commonwealth of
Independent States were either market or resource-seeking (Lankes
& Venables, 1998; Meyer, 1998).

Morisset (2000) used panel data of 29 African countries
over the period 1990-1997 to study the main determinants of FDI
in Africa. Morisset"s study stipulated that GDP growth rate and
openness are positively related to FDI in Africa. Campos and
Kinoshita (2003) used panel data of 25 transition economies from
1990-1998 to study the main determinants of FDI. They found that
natural resource endowment is the main determinant of FDI flows
in Africa.

Return on investment is a critical factor for rational
investors. Foreign direct investors will go to countries that pay
a higher return on capital. According to Asiedu (2001), measuring
the rate of return in developing countries is a difficult
process. The reason is that the capital market in developing
countries is under develop. The measurement which has been
adopted for this variable is to use the inverse of real GDP per
capita to measure the return on capital. The empirical result of
the relationship between real GDP per capita and FDI is diverse.
Edwards (1990) and Jaspersen et al. (2000) used the inverse of
income per capita as a proxy for the return on capital and they
concluded that real GDP per capita and FDI inflows as a ratio of
GDP are negatively related. Schneider and Frey (1985) and Tsai
(1994) studies found a positive relationship between the two
variables. This is based on the argument that a higher GDP per
capita implies better prospects for FDI in the host
country.

One of the ways in which FDI positively affects economic
growth is through the release of binding constraint on domestic
savings in the economy. FDI makes it through the process of
Capital accumulation. Given that domestic savings in Africa are
very low, which may result in a low investment rate and hence
lead to sluggish economic development (Ajayi, 2006). A study
conducted by Khan & Bamou (2006) on some selected developing
countries. They found that FDI has a positive effect on domestic
savings through foreign savings. Other studies also found results
that are consistent with those of Khan and Bamou (Asante, 2006,
Abedian, 2003). Less emphasize has been put on trying to see
effect of an improving domestic savings on FDI.

3.3 Conclusion

The literature findings have shown the great importance
of FDI in Africa. It has also highlighted that Africa needs
economic development, which is evidenced in the performance of
some macroeconomic indicators identified by the literature and
empirical studies. There has been an inconclusive finding on the
effect of some determinants of FDI on attracting FDI inflows to a
country, and few papers have undertaken a comparative study using
a linguistic approach. However, this paper has contributed to the
plethora of literature by adopting a linguistic approach to
identify the effect of some determinants of FDI inflows in
Arabophones and Francophones" countries.

Methodology

4.1 Types and sources of data

Secondary annual data were used in the study. The data
were sourced from the United Nations Conference on Trade and
Development (UNCTAD) data base and World Bank Development
indicators (WDI) database. The period considered for the study
was 1980-2010.

4.2 Model specification

The paper used the following variables in the model
specification to assess the effect of economic growth (EG),
inflation (INF), financial development (FD), openness (OPN),
balance of payments (BOP), domestic savings (DS), infrastructure
development (INFRD), government size (GS), return on investment
(RI), domestic investment (DI), and country specific dummies on
Foreign Direct Investment inflows in Arabophones and Francophone
non-landlocked Africa countries.

Table 2: Description of variables

The dynamic ordinary least square estimates for
heterogeneous panel

4.3 Techniques for data analysis

The paper first tests for Stationarity of the variables
included in the models. The variables that are non-stationary at
level were not tested for the existence of a co-integrating
relationship because the dependent variable (FDI inflows) was
stationary at level. Multi-Co linearity check was conducted using
correlation matrix. The paper also had undergone some diagnostic
tests such as Wald- test and Hausman test. Before estimating
panel DOLS, the Hausman test preferred a fixed specification to
random effects. The next step was to use a Dynamic ordinary least
square (DOLS) regression to estimate the effect of some FDI
determinants on FDI inflows in Arabophones and Francophone"s
countries on a comparative static analysis.

Empirical results
and interpretations

5.1 Panel Unit Root Test

The panel unit root tests were carried out to test for
the Stationarity of the variables used in the model
specification. The tests are necessary in order to avoid spurious
results. This study adopted two types of first
generation[2]panel unit root test such as Levin,
Lin and Chu (LLC) (2002) and Fisher-Type test using Augmented
Dickey Fuller ADF (Maddala and Wu, (1999)). The reasons why the
study has chosen LLC and ADF over Im, Pesaran and Shin (IPS)
(2003) and others is that; LLC generalize the Quah"s model which
allows for heterogeneity of individual deterministic effects
(constant and/or linear time trend) and heterogeneous serial
correlation structure of the error terms assuming homogeneous
first order autoregressive parameters. They assume that both
N and T tend to infinity but T
increase at a faster rate, such that N/T
approaches to zero. ADF is similar to IPS but both LLC and
Fisher- ADF allows for an unbalanced data in the test while IPS
does not allow for unbalanced data. Therefore, the tests are
suitable for this paper because its analysis was based on
unbalanced data specifications.

Levin, Lin and Chu (2002) as well as ADF- Fisher
Chi-square were proposed by Madala and Wu (1999) and they both
have the same null hypothesis of unit root is presence against
its alternatives.

In table 3 below, eight variables out of the eleven
variables used in this paper were stationary at levels including
the dependent variable FDI. Although the rest of the variables
were stationary after first differencing, the co – integration
test was not conducted because the dependent variables were
stationary at level. The two adopted tests such as an LLC and
ADF-fishers chi-square were consistent with their inferences. The
variables with asterisk indicated the Stationarity integral
order.

Table 3: Panel Unit
root[3]for Non-landlocked Africa
Countries

Probabilities are computed assuming
asymptotic normality. Critical values 1% * 5%**and Critical
values at 10%*** ASL represented Already Stationary at
Level

In table 4, Stationarity test was conducted by dividing
the non-locked countries into Arabophones and Francophone
countries in Africa. The table shows that four variables out of
the eleven variables were stationary at levels including the
dependent variable (FDI) and the rest became stationary after
first differencing using ADF-Fishers Chi-square test. The result
was partly consistent with intercept only and intercept and trend
specification. The inconsistency was that two additional
variables became stationary at levels with intercept and trend in
Arabophones. These included the dependent variable and trade
openness (OPN). In the Francophone specification, seven variables
out of the eleven variables became stationary at levels excluding
the dependent variables when included only intercept. The nature
of the variables showed the same results when intercept and trend
were included, although it differs with domestic investment
(DI).

Table 4: Panel Unit root for Comparative
analysis using ADF-Fishers Chi-square

Probabilities are computed assuming
asymptotic normality. Critical values 1% * 5%**and Critical
values at 10%*** ASL represented Already Stationary at Level. The
value in () represented variable that became stationary after
differencing and it is exclusively for stationary
test.

Although that the dynamic ordinary least squares (DOLS)
model adopted included the long run and short run effect but in
the advert of non-significant lags and lead variable the long run
relationship disappears. Therefore, panel Co-integration became a
necessary test for Arabophones and Francophone models.

5.2 Panel Co-integration
Test

This paper employed Pedroni (Engle-Ganger based) Panel
co-integration test suggested by Pedroni (1995, 1999, 2000).
These tests extend the Engle and Granger (1987) two-step strategy
to panels and relied on the ADF and PP principles.

Pedroni (1995, 1999, and 2000) proposed seven test
statistics for co-integration in a panel framework. Four of the
statistics are called panel co-integration statistics, which
pooled within-dimension based statistics (Pedroni, 1995, 1999).
The other three statistics developed by Pedroni (2000), are
called Group-mean Panel Co-integration statistics, which are
between-dimension panel statistics. The seven statistic is given
as:

Each of the panel test statistics will be distributed
asymptotically as a normal distribution.

The null hypothesis of no co-integration against the
alternative of co-integration was tested using the seven
statistics. According to the test if more than half of statistic
were statistical significant, then the null hypothesis of no
co-integration is rejected, otherwise you fail to reject the null
co-integration.

Table 5: Pedroni (Engle-Granger Based) Panel
Co-integration Test

Probabilities are computed assuming asymptotic
normality. Critical values 1% * 5%**and Critical values at 10%***
represents the level of significance of the
statistic.

In table 5, the co-integration model for Francophones
countries includes; Foreign Direct Investment, trade openness,
infrastructure development and financial development variables.
The co-integration model for Arabophones countries includes
foreign direct investment, trade openness, inflation, domestic
investment, financial development and balance of payment.
According to the test, there was a long run relationship between
the independent variables and foreign direct investment in both
model specifications. The variables included in the
Co-integration test were those variables which became stationary
after first differencing.

5.3: Model specification
test

Table 6: Correlation
Matrix[4]

The values without a bracket, [ ] and (
) denotes the correlation matrices for non-landlocked Africa,
Arabophones and Francophones countries
respectively.

Table 6 shows the correlation matrices of different
variables in the three models. This is one way of testing the
presence of multicollinearity in model estimation. According to
the test, none of the independent variable is highly correlated
to each other. The next specification test conducted by this
paper was Hausman test.

Table 7: Correlated Random effects-Hausman
test[5]

According to table 6, the proper model specification for
all the models was under fixed effect estimation which the paper
adopted.

5.4 Empirical Results and Discussions

This paper used Dynamic Ordinary Least Square (DOLS) to
answer all the objectives. The DOLS allowed for the inclusion of
both variables that were stationary at levels and those that
became stationary after differencing in the model. The lag and
lead of explanatory variables that were not stationary at level
were included in the model after differencing and insignificant
lag and lead variables were excluded in the final presentation of
the tables below.

Table 8 below shows empirical findings of the
determinants of FDI on FDI inflows in Arabophones, Francophone
and non-landlocked Africa countries. The non-landlocked countries
exclusively include only those countries that their official
language is either French in the case of Francophone or Arabic
for Arabophones. The model included in its panel estimation,
seven Francophone countries such as Gabon, Benin, Cote d"Ivoire,
Togo, Guinea, Madagascar, and Senegal.

Table 8: Linguistic approach effect of
FDI determinants in non-landlocked Africa
countries

Probabilities are computed assuming
asymptotic normality.

The five Arabophones countries included in the analysis
were Algeria, Egypt, Morocco, Tunisia, and Mauritania. The
variables included in the model were Investment (DI), Domestic
Savings (DS), Economic Growth (EG), Inflation (INF), and
Infrastructure Development (INFRD), Government Size (GS),
Openness (OPN), Balance of Payment (BOP), Financial Development
(FD), and Return on Investment. Durbin-Watson statistic of 1.709,
1.051, 0.977, 0.978 and 0.978 as in Arabophones, Francophone,
non-landlocked, inclusion of Arab dummy, and the inclusion of
Franc dummy respectively. Although all the models except
Arabophones model specification shows a presence serial
correlation but the F-statistic entailed that all the model
specification are good. That is, one can make inferences from the
result because the F-statistic is significant. In addition, the
R-squared and the Wald test shows that the model fitness was good
as well as the null hypothesis of overestimation were rejected in
the entire models. All the estimations were carried out under
fixed effect specification of unbalanced panel data suggested by
the Hausman test.

The result showed that the contemporaneous Financial
Development (FD), Domestic Savings (DS), Balance of Payment
(BOP), Trade Openness (OPN), lead of Domestic Investment (DI),
lead of Domestic Savings (DS) and lead of Infrastructure
Development (INFRD) are determinants of Foreign Direct Investment
(FDI) inflows in Arabophones countries that were statistically
significant. The same variables were identified as some
determinants of FDI inflows in Francophones countries except the
lead of Domestic Investment (DI), the lead of Domestic Savings
(DS), and lead of Infrastructure Development (INFRD) and
inclusion of lagged INFRD and inclusion contemporaneous INFRD and
DI were also statistically significant. The non-landlocked Africa
countries model showed that the same variables were also the
determinants of FDI inflows inclusive of contemporaneous Real
Economic growth (EG), contemporaneous DS, contemporaneous
Inflation (INF) and excluding their lagged and lead variables.
There was a slight change when Linguistic approach differences
were captured as dummy variables. Although, the dummies were not
statistical significant but it improves the model by the
identifying lead OPN as one of the determinants of FDI inflows in
non- Landlocked countries of Africa.

According to the result, an increase in the financial
development by a unit increases the FDI inflows to Arabophones
countries by 0.0003 units, in the non-landlocked by the same
units but reduces the FDI inflows to Francophone countries by
0.001 units. The effect was statistically significant. The result
is consistent with Hermes and Lensink (2003) and Nabamita and
Sanjukta (2008), which found a positive relationship between
financial development and foreign direct investment. However, the
negative effect with the Francophones countries can be associated
with the type of foreign direct investor"s motive.

Furthermore, an increase in infrastructure development
(INFRD) reduces the FDI inflows to Francophones countries by
0.0003 units and the effect is statistically significant.
Although the effect in both Arabophones and non-landlocked
countries have a positive effect but was not statistically
significant. Good infrastructures stimulate production, reduce
operating costs and invariably encourage FDI inflows (Wheeler
& Mody, 1992). It also increases the productivity of
investment and hence economic development. But this finding were
not consistent with the relationship between INFRD and FDI for
Francophones countries and that of Arabophones in the long-run.
According to the current paper's findings, an increase in INFRD
by a unit reduces FDI inflow to Francophones by 0.0003 units and
0.0008 units to Arabophones in the long-run.

There was a positive relationship between the domestic
investment (DI) and FDI inflows to Arabophones and Francophones
countries. This implies that FDI inflows crowd in DI in both
countries, and its effect on Francophones is statistically
significant while on Arabophones is not. However, domestic
investment has a positive effect on FDI inflows to Arabophones in
the long-run and statistically significant at 10 percent as
depicted from the empirical result.

Fiscal and monetary disciplines by the recipients"
government have a negative effect on the FDI inflows in their
country. These disciplines can be seen from the nature of their
balance of payment (BOP). The empirical findings in this paper
showed that an increase in BOP by a unit reduces the FDI inflows
by 0.0006 units, 0.001 units, and 0.0009 units in Arabophones,
Francophones and Non-landlocked Africa countries excluding
Lusophone countries of Africa respectively.

Another determinant of FDI inflows to Arabophones and
Francophones countries was trade openness. In theory, trade
openness effect on FDI inflows to an economy depends on the
investor"s motivation for engaging in FDI activities (Brainard,
1997; Markusen and Maskus, 2002; Navaretti and Venables, 2004).
According to the authors, a more open an economy becomes the
better for non-market seeking investors who would like to use the
destination as an export base. On the contrary, market seeking
investor whose target market is the host countries would prefer
less openness. This paper result to a certain extent is
consistent with the theory based on the non-market foreign direct
investor"s motive, that is, an increase in the trade openness
(OPN) by a unit increases the FDI inflows to Arabophones
countries 0.0412 units and to Francophones countries by
approximately the same units (0.0411 units). It also increases
the FDI inflows to Non-landlocked countries of Africa excluding
Lusophone countries by 0.0266 units and to the model that
captured the Linguistic differences as dummies in the short-run
by 0.030 units.

Contemporaneous inflation (INF), real economic growth
(EG) and domestic savings (DS) have a positive effect on FDI
inflows to Arabophones, Francophones and Non-landlocked excluding
Lusophone countries and the models with dummies. Although, the
positive effect is statistically significant on the other models
excluding Arabophones and Francophones models, the DS has a
negative effect on FDI inflows to Arabophones countries. That is
a percent increase of DS reduces FDI inflows to Arabophones
countries by 0.0004 units in the long run.