Cronfa - Swansea University Open Access Repository

_____________________________________________________________

This is an author produced version of a paper published in:

Journal of Behavioral and Experimental Finance

Cronfa URL for this paper:

http://cronfa.swan.ac.uk/Record/cronfa51109
_____________________________________________________________

Paper:
Abdel-Latif, H. (2019). FDI response to political shocks: What can the Arab Spring tell us?. Journal of Behavioral and
Experimental Finance
http://dx.doi.org/10.1016/j.jbef.2019.07.005

Released under the terms of a Creative Commons Attribution Non-Commercial No Derivatives License (CC-BY-NC-
ND).

_____________________________________________________________

This item is brought to you by Swansea University. Any person downloading material is agreeing to abide by the terms
of the repository licence. Copies of full text items may be used or reproduced in any format or medium, without prior
permission for personal research or study, educational or non-commercial purposes only. The copyright for any work
remains with the original author unless otherwise specified. The full-text must not be sold in any format or medium
without the formal permission of the copyright holder.

Permission for multiple reproductions should be obtained from the original author.

Authors are personally responsible for adhering to copyright and publisher restrictions when uploading content to the
repository.

http://www.swansea.ac.uk/library/researchsupport/ris-support/
 FDI response to political shocks:
What can the Arab Spring tell us?

Abstract
This paper examines the FDI response to political shocks. I first investigate whether
political or institutional quality is an important determinant of FDI inflows using a panel
VAR model in a dataset of 146 countries over 1989-2015. Then, I exploit the Arab Spring
incidence to measure the short-run effects of political shocks on FDI flows using the
differences-in-differences (DiD) estimator for a sub-sample of nineteen countries in the
MENA region. I account for possible bias of the DiD estimator resulting from dealing
with heterogeneous group of countries by using the propensity score matching based
on countries’ economic development and political settings. The findings show that a
positive shock to political quality would increase FDI flows which lends evidence to the
importance of political quality in determining FDI flows. In addition, I find that the Arab
spring has led to a drop in FDI flows to the MENA region.

Keywords: FDI, Political Shocks, MENA Region, Arab Spring, Panel VAR
JEL code: F21, D72, O47, C33

1. Introduction

Foreign direct investment (FDI) can be an important growth vehicle. 1 Economic theory
and empirical evidence explain this link via a number of transmission channels such
as know-how, propensity-to-export, human capital, and employment. 2 Policy makers,
especially in developing countries, put measures in place to ensure a favourable invest-
ment climate for foreign investors. While many countries were successful in attracting
considerable amounts of FDI, others failed to secure but modest levels of these flows.
Such observation has triggered empirical research on the determinates of FDI flows.
However, focusing primarily on economic factors, this research has paid relatively less
attention to political determinants of FDI. In fact, FDI might be even more sensitive to
political factors compared to economic ones; given that FDI mobility is an ex ante by

1 See, for example, Zhang (2001); Hansen and Rand (2006); Ayanwale (2007); Gursoy et al. (2013); Castro

(2014)
2 See Mankiw et al. (1992); Basu and Weil (1998); Borensztein et al. (1998); Galor and Weil (2000); Blomström

and Kokko (2002); Liu et al. (2002); Blomström and Kokko (2003); Javorcik (2004); Castro (2014); Goswami and
Haider (2014)
nature and illiquid ex post. That is, political risk appears to be a relevant determinant for
foreign investors3 .

Unsurprisingly, studies on political determinants of FDI has flourished recently; given

the unstable political arena in many countries around the world, especially in the MENA
region. To reign in foreign investment, perhaps offering financial and fiscal incentives
may not be sufficient. Political stability is a crucial factor in play for investors to ensure
the sustainability of these incentives in order to seize their advantages. Politically un-
stable countries with poor institutional settings are more likely to fail in attracting FDI
flows. This lends importance to research examining how the quality and stability of
political institutions in host countries would affect investors’ perception of risk levels
they will have to take on. Current studies examining the FDI-politics nexus focus on
long-run relationships, which overlooks the impact of sudden political shocks on the FDI
allocation behaviour. In this regard, the ‘Arab Spring’ has put forward such a question on
the research agenda for the coming years, which motivates this study. I first investigate
the relationship between FDI flows and political quality. Then, I exploit the incidence of
the Arab Spring to study how sudden political shocks may affect the FDI geographical
distribution, focusing on the short run effects of the Arab Spring on FDI flows to the
MENA region.

This is the first paper to study the effects of the Arab Spring on FDI flows. Thus, it
contributes to the body of literature that focuses on the political determinants of FDI. It
is extremely important for policy makers to understand how FDI flows may respond to a
political shock in a given country and/or region. Therefore, the findings are of direct
interest to policy makers who wish to evaluate the effects of sudden political shocks in
distorting FDI flows.

This paper proceeds as follows. Section 2 overviews the FDI flows and its relationship
with host countries political and economics characteristics, especially in the MENA
region. Section 3 reviews the existing literature. Section 4 explains the methodology. Sec-
tion 5 summarises the dataset and variables of interest. Section 6 presents the empirical
results. Finally, section 7 concludes.

2. Background

Considering the FDI flows worldwide, it has witnessed a tremendous increase in 2000s;
compared to 1980s and 1990s. Fig. 1 shows the evolution of FDI flows worldwide be-
tween 1989-2015. Although there is an obvious upward trend from the beginning of the
1990s, the figure shows that FDI flows have shown a distinctive increasing trend in the
2000s. More particularly, the average FDI flows rose from about $296 billion between
1989-1999 to about $1.2 trillion between 2000-2007. However, with the eruption of the
2007-2008 global financial crisis the FDI flows have dropped sharply. Two years after the

3 Although FDI is a more stable carrier of investment and less footloose when compared to other sources

of finance, empirical studies and experience have proven that international investors are quite sensitive to
political risks.

2
crisis, the FDI shows signs of recovery though never reached its before crisis peak. Yet,
the FDI flows remain above their 1990s level.

Figure 1: FDI flows (1989-2015), World

Looking at the regional distribution of the FDI flows, one can easily notice that Europe
has secured the lion’s share amongst the world’s regions, over the precedent decade.
Fig. 2, which plots FDI flows between 2005-2015 by region, shows that Sub-Saharan
Africa and MENA regions have always received a modest share of the FDI cake. More
specifically, while Europe has received, on average, $0.8 trillion Sub-Saharan African
and MENA countries only received $33 billion and $83 billion, respectively. The same
figures were $215 billion & $433 billion in Latin America and East Asia. Nevertheless,
as noted earlier, the global financial crisis of 2007-2008 seems to have negative impacts
on FDI flows in almost all regions, see Fig. 2. However, it is evident how uneven the
regional distribution of FDI is among different parts in the world economy. That is, the
MENA region seems to be running far behind in terms of attracting FDI flows. Thus, it is
important for policy makers in Sub-Saharan Africa and MENA countries to understand
what determines the FDI inflows.

This study focuses on how FDI responds to sudden political shocks in the MENA region.
What makes the MENA region a very interesting context for this study is the eruption of
the Arab Spring; a series of political shocks that might have knocked foreign investors’
confidence in the region. Although the MENA region seems to have benefited from the
surge in the world FDI flows, its share compared to other regions remains modest, Fig. 3.
In addition, like other regions, the MENA countries were not immune to the negative
effects of the global financial crisis on FDI flows. In Fig 3, which shows the evolution
of FDI flows to the MENA region between 2000 and 2015, although FDI showed some
signs of recovery after the global crisis in other regions, it witnessed another drop in the
MENA region from 2011 and onward, i.e., coincidence with the Arab Spring event.

3
 Figure 2: FDI flows (2005-2015), by region

Figure 3: FDI flows (2005-2015), MENA region

Since I focus primarily on the MENA region, Fig. 4 and Fig. 5 plot the average FDI
flows in the region between 2005-2015 against individual countries specific institutional
settings and economic development characteristics. For this purpose, I use the polty
index to proxy for country specific institutional quality and GDP growth rate to represent
economic development process in the host country4 . As can be seen in Fig. 4, although
Lebanon and Israel have the highest values for the polty index, yet they attract different
amounts of FDI flows, measured as a percentage of GDP. Nonetheless, both figures, 4
and 5, suggest that both institutional quality and economic development might be of
important determinants of the FDI flows into MENA countries.

4 More details about the dataset can be found in section 5.

4
 Figure 4: FDI & Polty (2005-2015), MENA countries

Figure 5: FDI flows & GDP growth (2005-2015), MENA countries

3. Literature Review

The uneven distribution of the FDI flows among different regions/countries stimulated
empirical research studying what determines FDI. Numerous studies have primarily
focused on the economic determinants of FDI. Such factors can be broadly classified into
two groups: pull factors and push factors. The ‘pull’ factors include host country-specific
attributes such as market size, per capita growth, good infrastructure, and trade open-
ness.5 The ’push’ factors are related to the conditions in the source country (Mathur and
Singh, 2013). However, the economic determinants of FDI flows have been extensively

5 See, Chen (1996); Lipsey (2000); Asiedu (2002); Durham (2004); Sekkat and Veganzones-Varoudakis (2007);

De Vita and Kyaw (2008), among others.

5
studied6 . Only few studies lend importance to the effect of political factors on FDI inward
flows. This body of literature has shown that FDI flows are sensitive to political factors
since that it represents a long-term investment decision and investors, therefore, have to
undertake a substantial amount of risk when choosing a destination.

In this context, overseas capital movements face different types of political risk such as
the risk of expropriation. For example, the theoretical model proposed by Asiedu et al.
(2009) shows that higher levels of expropriation risk would lead underinvestment and
would attenuate the optimal level of FDI. Thus, any unexpected government intervention
in the host country would be a typical example of political risk facing foreign investors
when deciding on a destination for their investment (Madani and Nobakht, 2014).

Nevertheless, the existing literature identifies three scenarios when explaining the effects
of political institutions and regime type on FDI. The first scenario implies that demo-
cratically elected regimes, in which institutions are fundamentally defined in favour of
free markets, reduce political risks facing businesses. For example, Madani and Nobakht
(2014) use a dynamic panel model of 31 upper middle income countries to find that
democracy enhances FDI inflows. Similar results can be found in Harms and Ursprung
(2002) who find that FDI tends to flow into countries with civil and political freedom.
Busse (2003) uses a panel data model to conclude a clear causation between democratic
rights and FDI per capita.

The second scenario in the literature asserts no link between democracy and FDI inflows.
Goswami and Haider (2014) refute that government failure can explain poor FDI inflow.
Similar results appear in Li (2009) and Schulz (2009), where they find no linkage between
FDI and political regime. Utilising the democracy-dictatorship (0,1) dataset of Cheibub
et al. (2010) and a dynamic panel data model, Castro (2014) find no evidence that relates
FDI to democracy. In addition, Mathur and Singh (2013) show that economic freedom in
the host country is what influences foreign investors’ decisions regardless of its political
freedom. The authors argue that more democratic countries with poor economic freedom
may receive less FDI flows.

The third scenario points to a ‘cosy’ relationship between foreign investors and autocrats
to the extent to conclude a negative impact of democracy on FDI flows. Their argument
rests on the freedom and the lack of accountability authoritarian rulers enjoy which
entitle them to offer relatively more generous incentives to foreign investors (Alesina and
Dollar, 2000). This scenario can be found in studies such as Haggard (1990); Busse (2004);
Adam and Filippaios (2007) who actually state that FDI may benefit from dictatorships.
Finally, Li and Resnick (2003) explains this argument further. The authors suggest that
democratic institutions are more likely to adopt policies that reflect popular sentiment,
which may not be in favour of foreign investors. Autocratic leaders, on the other hand,
adopt biased policies that favour the narrow elite control which are more likely to favour
foreign investors and facilitate collusion-based expansions in their market share.

6 A non-exhaustive list includes Balasubramanyam et al. (1996); Moran (1998); Noorbakhsh et al. (2001);

Moosa et al. (2002); Blonigen (2005); Helpman (2006).

6
In whichever case, research on the FDI impacts of political shocks is very scarce. For
example, empirical evidence reported in Busse and Hefeker (2007) point to a number
of political factors in the host country, such as the stability of the central government
stability, that are important in determining the level of FDI flows. Similar examples in-
clude Dutta and Roy (2011) and Roe and Siegel (2007). The authors argue that politically
unstable governments are more likely to adopt a more myopic attitude toward foreign
investors. In addition, Stasavage (2002), study how political uncertainty can deter private
investment across countries though the channel of risk perception. Finally, Janeba (2002)
shows how poor levels of government ‘credibility’ in the host country can explain why
developing countries seem not to be attractive destinations for foreign capital.

To sum up, scholars usually focus on exploring the long-run impact of a given set of
macroeconomic and/or political factors on the behaviour of foreign investors, in terms
of their country of preference. However, the results on the linkage between FDI inflows
and the quality of political institutions and regime type have remained inconclusive.
Meanwhile, very few studies looked at the short-run responses of FDI to sudden political
turmoil, which motivates the current study to fill several loopholes that currently exist in
the literature. More particularly, this study makes an attempt to answer two research
questions, which not yet are adequately covered in the literature: i) how do FDI flows
respond to political shocks?; ii) how did the Arab Spring, defined as a sudden political or
institutional shock, impact FDI inflows to the MENA region?

4. Methodology
I first employ a panel VAR modelling approach to investigate the FDI response to political
shocks. A simple VAR model in which all variables are endogenous and interdependent
can be presented as follows. Let Yt be a G ⇥ 1 vector of endogenous variables. Then, a
VAR representation of Yt can be presented as follows.

Yt = A0 (t) + A(l)Yt 1 + ut (1)

where ut ⇠ iid(0, ⌃u ). A panel VAR model can have the same structure above, how-
ever will encompass a cross sectional dimension in addition to its time representation.
Suppose that Yt = (y01t , y02t , . . . , y0Nt )0 , it follows that the panel formation of Eq. 1 can be
presented as follows.

Yit = A0i (t) + Ai (l)Yt 1 + uit (2)

where i = 1, . . . , N & t = 1, . . . , T and uit = G ⇥ 1 is a vector of random disturbances. In
this study, I estimate a trivariate model that includes FDI inflows, a proxy for political
settings and a measure of development process in the host country. After estimating
the panel VAR model, as well as confirming its stability, I produce the impulse response
functions in order to examine how FDI flows respond to political shocks in the whole
sample.

In addition, to capture the possible effect of the Arab Spring on FDI flows through the
institutional quality channel, I employ the Differences-in-Differences (DiD) estimation.
7
The DiD strategy compares countries that are from the MENA region with those that are
not. With this purpose, it is necessary to define a treatment group composed of countries
in the MENA region which were subject to a political shock (i.e., the Arab Spring), and
a control group formed by countries which were not subject to the shock (i.e., outside
the MENA region). The DiD estimator will capture the differential effect of the shock on
the treatment group relative to non-treatment countries. To apply the DiD estimator all
that is necessary is to measure outcomes in the treatment and control groups both before
and after the shock. The simple DiD estimator compares the mean of the outcome in
treatment and control groups which is well justified on the grounds that they should not
have any systematic differences in any other pre-treatment variable. Let µit is the mean of
the outcome in group i in time t, where i = 0 if the country falls outside the MENA region
(control group) and i = 1 if the country belongs the the MENA region (treatment group).
Define t = 0 as a pre-treatment period (year < 2011) and t = 1 as post-treatment period
(year 2011). Thus, a simple DiD estimator can be expressed as (µ11 µ10 ) (µ01 µ00 ).
The first term is the change in outcome for the treatment group and the second term is
the change in outcome for the control group. A DiD regression based estimator can be
obtained by estimating the following equation.

yit = 0 + 1 MENAi + 2S pringt + 3 MENAi .S pringt + ✏it (3)

Where yit is the outcome of interest (FDI); MENAi is a dummy variable that takes the
value of one if the country belongs to MENA region (treatment group) and zero if the
country is not in the MENA (control group); S pringt is a dummy that takes the value
of one for year 2011 (post-treatment) and zero for year < 2011 (pre-treatment) and
MENAi .S pringt is the interaction term of the previous dummies which is just a dummy
variable that takes the value of one only for the treatment group in the post-treatment
period. The DiD estimator is the OLS estimator of 3 , the coefficient of the interaction
term between MENAi and S pringt . The next step is to include additional controls in the
estimation, in order to control for observable variables that could affect the outcome of
interest (i.e., FDI inflows). Therefore, equation 3 can be written as follows.

yit = 0 + 1 MENAi + 2S pringt + 3 MENAi .S pringt + Xi + ✏it (4)

Where Xit can be a series of control variables related to economic and political determi-
nants of FDI flows.

The DiD estimator is expected to be an unbiased estimate of the change in FDI inflows
due to a political shock (i.e, the Arab Spring) if both groups were similar except for being
subject to the treatment (i.e., the shock). However, this is a very strong assumption as
countries are expected to be of different characteristics and contexts. To account for this
bias, I utilise the propensity score matching (PSM) technique to estimate the average
treatment effect of the treated (ATT). The PSM methodology identifies the average
treatment effect by comparing the outcome of MENA countries and other countries
which, a priori, have similar characteristics. Therefore, the PSM is expected to reduce the
bias generated by any unobservable confounding factors. In this context, the propensity
score can be defined as the conditional probability of the incidence of a political shock
given pretreatment characteristics.

8
 p(X) ⌘ Pr(D = 1|X) = E(D|X) (5)
where D = {0, 1} is the indicator of exposure to a political shock and X is the multidimen-
sional vector of pretreatment host country specific characteristics. If the propensity score
p(Xi ) for country i is known, then the ATT can be estimated as follows:

⌧ ⌘ E{Y1i Y0i |Di = 1}

= E[E{Y1i Y0i |Di = 1, p(Xi )}]
= E[E{Y1i |Di = 1, p(Xi )} E{Y0i |Di = 0, p(Xi )}|Di = 1]
where the outer expectation is over the distribution of (p(Xi )|Di = 1) and Y1i and Y0i are
the potential outcomes in the two counterfactual situations of receiving a political shock
and not receiving a political shock. However, this requires two conditions: (i) balancing
of the pretreatment variables given the propensity score. If p(X) is the propensity score,
then
D ? X|p(X)
and (ii) unconfoundedness given the propensity score. Suppose that assignment to
treatment (a political shock) is unconfounded, i.e., Y1 , Y0 ? D|X, then assignment to
treatment is unconfounded given the propensity score, i.e., Y1 , Y0 ? D|p(X). If the
balancing condition is met, observations with the same propensity score must have the
same distribution of observable (and unobservable) characteristics independently of
being subject to a political shock. In other words, for a given propensity score, receiving
a political shock is random and therefore treated and control units should be on average
observationally similar. In this study, I employ a probit model to estimate the propensity
score based on country specific political settings and economics development processes.

5. Dataset

The dataset includes annual data from 1989 to 2015 for 146 countries including nineteen
MENA countries, see Table 6 in the appendix. The variables of interest are FDI inflow (%
GDP) and GDP annual growth rate. Both variables were extracted from the World Bank
WDI database. I use the polity index to proxy for country-specific political settings. The
polty index, which comes from Marshall et al. (2016) database, varies from -10 to 10. The
index is based on sub-scores for constraints on the chief executive, the competitiveness
of political participation, and the openness and competitiveness of executive recruitment.
Higher values denote more democratic institutions. Marshall et al. (2016) define a polity
within the range [6,10] as a coherent democracy, one in the range [-10,-6] as a coherent
autocracy, and one in the range [-5,5] as an incoherent regime. Formally, it is computed
as the difference between a democracy index and an autocracy index, each ranging from
0 to 10. See Table 1 for summary statistics for the dataset.

6. Empirical Results

To estimate the panel VAR model, Eq. 2, I follow the estimation routine suggested in
Abrigo et al. (2016), who build on the generalized method of moments (GMM) frame-

9
 Table 1: Descriptive Statistics - overall sample

Variable Mean Std. Dev. Min Max Observations

FDI overall 3.357 5.230 -45.392 40.630 N = 3942
between 4.044 -19.049 25.689 n = 146
within 3.334 -30.375 29.066 T = 27
Polty overall 3.055 6.668 -11.000 11.000 N = 3942
between 6.118 -10.000 10.000 n = 146
within 2.697 -12.575 13.870 T = 27
GDP overall 3.421 4.408 -34.188 52.388 N = 3942
between 3.004 -17.182 16.458 n = 146
within 3.234 -25.524 39.351 T = 27

work7 . The authors apply forward orthogonal deviation proposed by Arellano and
Bover (1995) to remedy for the weaknesses of the first-difference transformation when
estimating dynamic panel models8 . I use information criteria to select the optimal lag
order (i.e., in both panel VAR specification and moment condition). Based on the model
selection criteria, first-order panel VAR is the preferred model, since it has the smallest
value for the information criteria. Therefore, I fit a first-order panel VAR model using
GMM estimation, see Table 2 for the estimated coefficients. The results in Table 2 suggest
that FDI is positively related to institutional quality in the host country. More particularly,
the panel VAR estimation shows that a one point increase in the polty index (i.e., indicat-
ing improvement in institutional quality in the host country) increases the FDI inflows
(measured as a percentage of GDP) by 3.4%. The estimated coefficient is statistically
different from zero at the one percent significance level. In order to estimate the FDI
response to political shocks, I exploit the panel VAR set up to produce the impulse
response functions (IRFs). However, it is important to confirm the stability of the model.

Table 2: Panel VAR estimation

Eq.
Coeff. FDI Polty GDP
L.FDI 0.981*** 0.01* -0.004**
L.Polty 0.034*** 0.853*** 0.028***
L.GDP -0.004** -0.017** 0.961***

The stability of the panel VAR model implies that it is invertible and has an infinite-
order vector moving-average (VMA) representation. This would ensure a meaningful
interpretation to the estimated IRFs and forecast error variance decompositions (FEVD).
I confirm the eigenvalues stability condition, where all accompanied values lie within
the unit circle. In addition to the stability condition of the estimation, I could employ
the panel VAR estimation to investigate the Granger causality between the underlying

7 The GMM estimator has been found to perform fairly well, especially in fixed T and large N.
8 They subtract the average of all available future observations, thereby minimizing data loss.
10
variables included in the model. Results of the Granger causality tests, presented in
Table 3, show that institutional quality (i.e., polty) Granger-causes FDI at the one percent
confidence level.
Table 3: Granger causality tests

Eq.
Coeff. FDI Polty GDP
FDI NA 2.914* 5.231**
Polty 60.475*** NA 22.22***
GDP 4.853** 5.579** NA
ALL 70.388*** 7.764** 23.378***

To this end, since that panel VAR estimates are seldom interpreted independently, I
proceed to estimate the IRFs as I am interested in examining the impact of political
shocks on FDI within a system of equations. Although the simple IRFs have no causal
interpretation, a shock on one variable is likely to be accompanied by shocks in other
variables, as well, since the innovations ✏it are correlated contemporaneously. Fig. 6
presents the IRFs from the panel VAR model for all variables in the system along with its
confidence bands. The IRF confidence intervals are estimated using Gaussian approxi-
mation based on 200 Monte Carlo draws from the estimated panel VAR model 9 .

9 Estimates of the FEVD along with its standard errors and confidence intervals are available but not shown

here to save on space.

11
 Figure 6: Impulse response functions IRFs from panel VAR model

As shown by Fig. 6, a one standard deviation (1SD) shock to the polty index would
lead to an increase in the FDI inflows by 0.2% of the GDP. This response seems to be
statistically significant at the five percent significance level. This implies that institutional
quality in the host country is an important determinant of FDI flows. In addition, the
IRFs show that institutional quality matters for economic development processes as well.
More particularly, a 1SD shock to the polty index seems to have a positive impact on GDP
growth, which is equivalent to about 0.1%. Again, this effect of the change of polty index
on GDP growth is statistically different from zero at the five percent significance level.
Thus, the results here suggest that institutional quality promotes growth and enhances
FDI flows.

The second objective in this paper is to estimate the short-run effects of the Arab Spring
on FDI inflows to the MENA region using the DiD estimator described in section 4. Table
4 presents the simple DiD estimation with no additional controls. Results in Table 4
provide the mean values of the outcome variable (i.e., FDI) for both control and treatment
groups, both prior to and post the Arab Spring periods. The DiD estimator is equal to
the difference across the two periods of the difference between treatment and control
groups. Differencing the mean values of the outcome variables between the two periods
for the treatment group gives the effect of the Arab Spring on FDI plus the effect of
any common shock such as the global financial crisis (or a time trend) that affect both
groups, MENA and non-MENA countries. The difference between periods for the control
group provides an estimate of such additional non-MENA related factors. Therefore,
differencing the estimate across periods for the treatment group with that of the control
12
group should offer an estimate of the effect of the Arab Spring on FDI inflows.

Table 4: Simple DiD estimation with no controls

Outcome var. FDI S. Err. t P>t

Baseline
Control 3.113
Treated 1.878
Diff (T-C) -1.236 0.27 -4.58 0.000***
Follow-up
Control 5.335
Treated 3.81
Diff (T-C) -1.525 0.566 -2.69 0.007***
Diff-in-Diff -0.289 0.627 -0.46 0.645

Defining the prior Arab Spring period as a baseline, the average FDI flows to the control
group (non-MENA countries) were 3.113% while in the treatment group this figure was
1.878%. Moving to the post Arab Spring period (T 2011), it is noted that both groups
had increased levels of FDI inflows. Specifically, FDI inflows have increased by 5.335 and
3.81 percentage points in the control and treatment groups, respectively. Thus, the simple
DiD estimator is equal to 0.289 percentage points. This implies that, on average, FDI
flows have responded to the shock (Arab Spring) by decreasing their flows to the MENA
region. However, the estimated DiD coefficient seems to be statistically insignificant.

Given the heterogeneity existing among countries in both groups (control and treatment),
the DiD estimation is likely to be biased. To account for this bias, I employ the propensity
score matching (PSM), thereby maximising the observable similarity between treatment
and control groups. As an alternative to linear regression, the PSM analysis allows us to
create the two groups that have similar characteristics so that a comparison can be made
within these matched groups. Implementation of the PSM methodology follows the two
step procedure whereby in the first step each country’s probability (propensity score)
of receiving a political shock is assessed conditional to a set of explanatory variables. I
include the polty index and GDP growth as controls within the first stage of the model to
ensure that the two groups are matched on similar characteristics: institutional quality
and economic development process. Consequently, the treatment and control group
countries are matched on the basis of their propensity scores. I present the results in
Table 5 based on the kernel matching method. The procedure involves taking each treated
country (MENA) and identifying non-treated countries (non-MENA countries) with
the most similar propensity scores. As the linear regression estimates in Table 4, the
matching results in Table 5 suggest that FDI flows into the MENA region have dropped,
on average, by 0.854% as a response to the shock after the Arab Spring. The estimated
DiD coefficient is statistically significant at the five percent level.

13
 Table 5: DiD estimation after matching

Outcome var. FDI S. Err. t P>t

Baseline
Control 2.581
Treated 1.947
Diff (T-C) -0.633 0.174 -3.64 0.000***
Follow-up
Control 5.298
Treated 3.81
Diff (T-C) -1.487 0.363 -4.1 0.000***
Diff-in-Diff -0.854 0.403 -2.12 0.034**

7. Conclusion

FDI flows are an important driving force of economic growth due to their positive exter-
nalities within the host country. Although FDI, as a long-term investment, may be able
to withstand crises more effectively, empirical evidence shows that FDI might be very
sensitive to political instability. Therefore, this study tried to assess the FDI response
to political shocks. To do so, the paper proceeded in two stages. First, I estimated a
panel VAR model for 146 countries over the period of 1989-2015. I used the impulse
response functions (IRFs) resulted from the panel VAR model to examine the effects of
political shocks on FDI flows in the overall sample. Second, drawing on the experience
of the MENA region, I exploited the incidence of the Arab Spring to estimate the short
run effects of sudden political shocks on the FDI flows. For this purpose, I employed
the differences-in-differences (DiD) framework. To account for possible DiD bias arises
from cross-country heterogeneities, I utilised the propensity score matching (PSM) to
compare the outcome of interest (i.e., FDI flows) in countries with similar polty scores
and economic growth rates.

To this end, the results shows that institutional quality in the host country is an impor-
tant determinant of FDI flows. This empirical evidence supports many studies in the
literature such as Harms and Ursprung (2002) and Busse (2003), while contradicts the
findings in other studies such as Li (2009) and Schulz (2009). In addition, I find evidence
that the Arab Spring, as a sudden political shock, has led to a drop in FDI flows to the
MENA region. The findings are of a significant importance to policy makers who wish to
evaluate the role of sudden political shocks in distorting FDI flows. Constantly erupting
political shocks tend to shake investors confidence, creating unnecessary turbulences to
the macroeconomic fundamentals, and negatively impacting development plans. And
FDI flows are not immune to those negative impacts. Therefore, policy makers should
design timely and appropriate policy responses to political shocks, given their negative
effect on FDI flows.

14
References
Abrigo, M. R., Love, I. et al. (2016), ‘Estimation of panel vector autoregression in stata’, Stata Journal 16(3), 778–
804.
Adam, A. and Filippaios, F. (2007), ‘Foreign direct investment and civil liberties: A new perspective’, European
Journal of Political Economy 23(4), 1038–1052.
Alesina, A. and Dollar, D. (2000), ‘Who gives foreign aid to whom and why?’, Journal of economic growth
5(1), 33–63.
Arellano, M. and Bover, O. (1995), ‘Another look at the instrumental variable estimation of error-components
models’, Journal of econometrics 68(1), 29–51.
Asiedu, E. (2002), ‘On the determinants of foreign direct investment to developing countries: is africa different?’,
World development 30(1), 107–119.
Asiedu, E., Jin, Y. and Nandwa, B. (2009), ‘Does foreign aid mitigate the adverse effect of expropriation risk on
foreign direct investment?’, Journal of International Economics 78(2), 268–275.
Ayanwale, A. B. (2007), ‘Fdi and economic growth: Evidence from nigeria’, African Economic Research Consortium
.
Balasubramanyam, V. N., Salisu, M. and Sapsford, D. (1996), ‘Foreign direct investment and growth in ep and
is countries’, The Economic Journal pp. 92–105.
Basu, S. and Weil, D. N. (1998), ‘Approperiate technology and growth’, Quarterly Journal of Economics 113(4).
Blomström, M. and Kokko, A. (2002), FDI and human capital: a research agenda, Citeseer.
Blomström, M. and Kokko, A. (2003), Human capital and inward FDI, Vol. 3762, Centre for Economic Policy
Research.
Blonigen, B. A. (2005), ‘A review of the empirical literature on fdi determinants’, Atlantic Economic Journal
33(4), 383–403.
Borensztein, E., De Gregorio, J. and Lee, J.-W. (1998), ‘How does foreign direct investment affect economic
growth?’, Journal of international Economics 45(1), 115–135.
Busse, M. (2003), Democracy and fdi, Technical report, HWWA discussion paper.
Busse, M. (2004), ‘Transnational corporations and repression of political rights and civil liberties: An empirical
analysis’, Kyklos 57(1), 45–65.
Busse, M. and Hefeker, C. (2007), ‘Political risk, institutions and foreign direct investment’, European journal of
political economy 23(2), 397–415.
Castro, D. (2014), Foreign Direct Investment and Democracy, Bryant University.
Cheibub, J. A., Gandhi, J. and Vreeland, J. R. (2010), ‘Democracy and dictatorship revisited’, Public Choice
143(1-2), 67–101.
Chen, C.-H. (1996), ‘Regional determinants of foreign direct investment in mainland china’, Journal of Economic
Studies 23(2), 18–30.
De Vita, G. and Kyaw, K. S. (2008), ‘Determinants of capital flows to developing countries: a structural var
analysis’, Journal of Economic Studies 35(4), 304–322.
Durham, J. B. (2004), ‘Absorptive capacity and the effects of foreign direct investment and equity foreign
portfolio investment on economic growth’, European economic review 48(2), 285–306.
Dutta, N. and Roy, S. (2011), ‘Foreign direct investment, financial development and political risks’, The Journal
of Developing Areas 44(2), 303–327.
Galor, O. and Weil, D. N. (2000), ‘Population, technology, and growth: From malthusian stagnation to the
demographic transition and beyond’, American economic review pp. 806–828.
Goswami, G. G. and Haider, S. (2014), ‘Does political risk deter fdi inflow?: An analytical approach using panel
data and factor analysis’, Journal of Economic Studies 41(2), 233–252.
Gursoy, F., Sekreter, A. and Kalyoncu, H. (2013), ‘Fdi and economic growth relationship based on cross-country
comparison’, International Journal of Economics and Financial Issues 3(2), 519–524.
Haggard, S. (1990), Authoritarianism and democracy: political institutions and economic growth revisited, Program on
US-Japan Relations, Harvard University.
Hansen, H. and Rand, J. (2006), ‘On the causal links between fdi and growth in developing countries’, The
World Economy 29(1), 21–41.
Harms, P. and Ursprung, H. W. (2002), ‘Do civil and political repression really boost foreign direct investments?’,
Economic Inquiry 40(4), 651–663.
Helpman, E. (2006), Trade, fdi, and the organization of firms, Technical report, National Bureau of Economic
Research.
Janeba, E. (2002), ‘Attracting fdi in a politically risky world’, International Economic Review 43(4), 1127–1155.
Javorcik, B. S. (2004), ‘Does foreign direct investment increase the productivity of domestic firms? in search of
spillovers through backward linkages’, American economic review pp. 605–627.
15
Li, Q. (2009), ‘Democracy, autocracy, and expropriation of foreign direct investment’, Comparative Political
Studies .
Li, Q. and Resnick, A. (2003), ‘Reversal of fortunes: Democratic institutions and foreign direct investment
inflows to developing countries’, International organization 57(01), 175–211.
Lipsey, R. E. (2000), ‘Inward fdi and economic growth in developing countries’, Transnational Corporations
9(1), 67–96.
Liu, X., Burridge, P. and Sinclair, P. J. (2002), ‘Relationships between economic growth, foreign direct investment
and trade: evidence from china’, Applied Economics 34(11), 1433–1440.
Madani, S. and Nobakht, M. (2014), ‘Political regimes and fdi inflows: Empirical evidence from upper middle
income countries’, Journal of Finance and Economics 2(3), 75–82.
Mankiw, N. G., Romer, D. and Weil, D. N. (1992), A contribution to the empirics of economic growth, Technical
report, National Bureau of Economic Research.
Marshall, G. M., Keith, J. and Robert, G. T. (2016), ‘Polity iv project: Political regime characteristics and
transitions, 1800-2009’, Dataset Users Manual. Center for International Development and Conflict Management,
University of Maryland .
Mathur, A. and Singh, K. (2013), ‘Foreign direct investment, corruption and democracy’, Applied Economics
45(8), 991–1002.
Moosa, I. A. et al. (2002), Foreign Direct Investment Theory, Evidence and Practice, Basingstoke, UK: Palgrave,.
Moran, T. H. (1998), Foreign direct investment and development: The new policy agenda for developing countries and
economies in transition, Peterson Institute.
Noorbakhsh, F., Paloni, A. and Youssef, A. (2001), ‘Human capital and fdi inflows to developing countries:
New empirical evidence’, World development 29(9), 1593–1610.
Roe, M. J. and Siegel, J. I. (2007), ‘Political instability and financial development’, Unpublished Manuscript,
Harvard Law School .
Schulz, H. (2009), ‘Political institutions and foreign direct investment in developing countries: Does the sector
matter?’, Available at SSRN 1403983 .
Sekkat, K. and Veganzones-Varoudakis, M.-A. (2007), ‘Openness, investment climate, and fdi in developing
countries’, Review of Development Economics 11(4), 607–620.
Stasavage, D. (2002), ‘Private investment and political institutions’, Economics & politics 14(1), 41–63.
Zhang, K. H. (2001), ‘Does foreign direct investment promote economic growth? evidence from east asia and
latin america’, Contemporary economic policy 19(2), 175–185.

16
 Appendix

Table 6: A list of countries in the dataset

Algeria* Czech Republic Israel* Nepal Sudan

Angola Denmark Italy Netherlands Suriname
Argentina Djibouti* Jamaica New Zealand Swaziland
Armenia Dominican Republic Japan Niger Sweden
Australia Ecuador Jordan* Nigeria Switzerland
Austria Egypt, Arab Rep.* Kazakhstan Norway Syrian Arab Republic*
Azerbaijan El Salvador Kenya Oman* Tajikistan
Bahrain* Eritrea Kuwait Pakistan Tanzania
Bangladesh Estonia Kyrgyz Republic Panama Thailand
Belarus Ethiopia Lao PDR Papua New Guinea Togo
Belgium Fiji Latvia Paraguay Trinidad and Tobago
Benin Finland Lebanon* Peru Tunisia*
Bhutan France Lesotho Philippines Turkey
Bolivia Gabon Liberia Poland Turkmenistan
Botswana Gambia, The Libya* Portugal Uganda
Brazil Germany Lithuania Qatar* Ukraine
Bulgaria Ghana Macedonia, FYR Romania United Arab Emirates*
Cambodia Greece Madagascar Russian Federation United Kingdom
Cameroon Guatemala Malawi Rwanda United States
Canada Guinea Malaysia Saudi Arabia* Uruguay
Central African Republic Guinea-Bissau Mali Senegal Uzbekistan
Chad Guyana Mauritania Sierra Leone Venezuela, RB
Chile Haiti Mauritius Singapore Vietnam
China Honduras Mexico Slovak Republic Yemen, Rep.*
Colombia Hungary Moldova Slovenia Zambia
Comoros India Mongolia Solomon Islands Zimbabwe
Congo, Rep. Indonesia Morocco* Somalia
Costa Rica Iran* Mozambique South Africa
Croatia Iraq* Myanmar Spain
Cyprus Ireland Namibia Sri Lanka
Note: * demotes a country in the MENA region.

17