Terrorism and its impact on inward FDI: an empirical analysis
on developed and developing countries.
Master Thesis University of Groningen Faculty of Economics and Business MSc International Economics and Business
January 8th, 2019
By Marlou Driessen
Student number: S2481847
Student mail: m.driessen.2@student.rug.nl
Supervisor: Dr. T. Kohl
1
Abstract
As the threat of terrorism has increased in the last few decades, the question on how terrorism influences international business and economic development has obtained more interest. In this paper, I will follow up on this question by analyzing how terrorism specifically affects inward FDI as it is a component of both international business and economic development. I extend this analysis to how it differs between developed and developing countries. I find that the intensity of terrorism could negatively impact inward FDI. Furthermore, I find that not all terrorist attacks affect FDI flows in a similar way, but that the target type matters. Terrorist attacks on diplomatic- and transportation-related targets negatively affects inward FDI flow in developed countries, whereas government-, military-, and tourist-related terrorist attacks negatively impact inward FDI flows in developing countries. Thus, in order to effectively attract more foreign capital, developing and developed countries should implement different counterterrorism mechanisms based on what type of terrorist attacks hurts their country.
2
Contents
1. Introduction ...3
2. Literature review ...5
2.1 Terrorism ...5
2.2 FDI and the effects on the host country ...6
2.3 FDI and its determinants ...8
3. Theory Development ...9 4. Methodology ...12 4.1 Variables ...12 4.2 Sample ...15 4.3 Model specifications ...16 5. Results ...18 5.1 Basic results ...18 5.2 Robustness check...22 6. Discussion ...24 6.1 Discussion of results ...24 6.2 Limitations ...25 6.3 Future research ...27 7. Conclusion ...27 8. References ...28
APPENDIX A. Target types and country lists ...32
APPENDIX B. Descriptive summary and correlation matrix ...34
APPENDIX C. Robustness check: FDI stock ...36
APPENDIX D. Robustness check: transnational terrorism ...38
3
1. Introduction
The idea of economic globalization due to an increase in international trade and international capital flows has been shaken up by an exogenous factor: terrorism. As illustrated in Figure 1, the number of terrorist attacks has increased worldwide since the turn of the century. The amount of casualties has seen an even larger rise; thereby, demonstrating that terrorism has become more destructive and lethal
(Czinkota, Knight, Liesch & Steen, 2010). Although terrorism often concentrates on developing
countries, terrorist attacks in Paris, Brussels, London, and Manchester in the last few years have shocked the developed world as well.
Not only does terrorism lead to loss of life, it also directly causes considerable damage to infrastructure, buildings and equipment. This loss of capital can cause for disruptions in (international) business activities and the economic development of a country. Thus, indirectly hurting the economy and the business environment.
Figure 1. Number of terrorist attacks and casualties (fatalities and injuries) caused by terrorism, 1970-2017.
Source: Constructed with data from the GTD (Global Terrorism Database, 2018)
Since the attacks on the World Trade Center in 2001, more research has been directed to terrorism and its economic consequences. This research can be divided into two topic areas: how terrorism impacts economic growth (Abadie & Gardeazabal, 2003; Meierrieks & Gries, 2012) and what the implications are of terrorism on international business activities, such as the tourism industry, international trade and
foreign direct investment (FDI).1
Notwithstanding the increase in interest into this topic, Czinkota et al. (2010) still pleads for more scholarly research on how the intense increase in terrorism affects international business. Due to globalization, economies are becoming more dependent on international economic ties, such as international trade and investment. Thus, to understand the long-term economic outcomes of terrorism,
4 we also need to comprehend how terrorism affects channels of international business (Czinkota et al., 2010).
One of these channels of international business that has seen an increased interest is FDI. In recent years, FDI flows have fluctuated. According to UNCTAD, FDI inflows fell with 23 percent in 2017 compared to 2016 (see Figure 2). In particular, developed countries saw a 37 percent decrease in FDI inflows relative to 2016 (UNCTAD, 2018). Interestingly, UNCTAD conducted a business survey in 2016 with executives on which factors or phenomena they expect to encourage or discourage FDI. 73% of the executives anticipated that terrorism would lead to a decrease in global FDI (UNCTAD, 2016).
Figure 2. FDI inflows in US dollars, global and by group of economies, 1970-2017.
Source: Constructed with data from the UNCTAD FDI/TNC database (UNCTAD, 2018).
This leads to the question whether the increase in terrorism is one of the reasons for the decline in FDI inflows. As countries’ economies have become more dependent on FDI, it is important to understand whether terrorism is detrimental to the FDI inflow of a country. The pioneers in this line of research are Enders and Sandler (1996) who found that in the period 1975 to 1991, terrorism had a negative effect of 13.5% on net foreign direct investment (NFDI) in Spain and a negative 11.9% on NFDI in Greece. This was followed by a handful of other studies that examined the impact of terrorism on FDI on samples of multiple countries (e.g. Abadie & Gardeazabal, 2008; Powers & Choi, 2012).
5 This ambiguity in results calls for more research to cover the gaps in the terrorism- FDI research area. According to Sandler and Enders (2008), research should distinguish between developing and developed countries as they have different economic and political institutions. They argue that countries with developed institutions are more likely to absorb terrorist attacks better; thus, not including the development level of a country could lead to biased results. Although studies on terrorism and FDI have used different samples in terms of including developed or developing or both type of countries, there are no studies that examined the effect of the development level on the relation between terrorism and FDI. It would be interesting to comprehend whether the development level moderates the effect of terrorism on FDI, since it could have implications on the effective measures to counter the economic consequences of terrorism in a country.
Moreover, previous research often utilize the number of terrorist attacks as their indicator for terrorism. They, therefore, group all the terrorist attacks together. According to Mancuso, Dirienzo and Das (2010), not all terrorist attacks will lead to a similar effect on FDI. Hence, they argue that the nature of a terrorist attack should be taken into account when analyzing the impact of terrorism on FDI. An attack on a power line or plant will more likely change the decision of a multinational to locate their FDI elsewhere than an attack on a local church, given that the latter will have less impact on the business activities of a multinational. Thus, in order to understand how terrorism affects FDI, it is necessary for research to distinguish between different types of targets of a terrorist attack (Powers & Choi, 2012).
Similarly to understanding the effect of the development level of a country, the effect of different targets of terrorism inward FDI of a country is valuable information for a government and policy-makers. This information can enhance the effectiveness of a government's policy to counter terrorism and its economic consequences. Therefore, the aim of this paper is to obtain a better understanding on how terrorism impacts inward FDI.
This paper will answer four questions. First, does terrorism impact FDI inflows into host countries? Second, does terrorism have a larger impact on FDI inflow into developing countries than in developed countries? Third, does the nature of a terrorist attack matters for whether terrorism affects FDI inflows? And does this effect differ between developing and developed countries?
The paper is organized as follows. In section 2 previous literature on terrorism and FDI is discussed. This literature is used to construct the hypotheses in section 3. Section 4 describes the data, variables and models that are used to analyze the hypotheses. Section 5 provides empirical evidence on the effect of terrorism on FDI. Section 6 and 7 conclude this paper by providing a short summary of results, its implications, this research limitations and suggestions for future research.
2. Literature review
2.1 Terrorism
6 Previous research on the relationship between terrorism and FDI often distinguishes between two types of terrorism: domestic and transnational. Enders et al. (2006) describe domestic terrorism as homegrown and home-directed with implications merely for the host country, whereas transnational terrorism happens in one country and involves victims, targets, perpetrators, and/or governments from another country.
Although terrorist attacks have shocked the world in the last few decades, there are only a handful of studies that examine the effect of terrorism on FDI worldwide. Abadie and Gardeazabal (2008) find in their study of 183 countries that in 2003-2004 that the risk of terrorism heavily influences FDI across countries. The authors argue that an increase in terrorist risk is associated with a decline in FDI position of about 5% of GDP. Bandyopadhyay et al. (2014) find that in their sample of 78 developing countries that both domestic as transnational terrorism depress FDI in the period of 1984 to 2008.
However, not every research finds a significant relationship between terrorism and FDI. Enders et al. (2006) find that from 1994 to 2004 transnational terrorism had a significant negative effect upon US FDI in OECD countries. However, this effect disappeared when the non-OECD countries were considered in the analysis. Li (2006) finds no significant evidence that either anticipated or unanticipated terrorism has a direct effect on FDI in his sample of 129 countries worldwide in the period 1976 to 1999.
Powers and Choi (2012) argue that not all terrorist attacks have the same detrimental effect on business opportunities. For instance, they argue that a multinational will only be cautious by investing into a business in a country or region that recently experienced terrorist attacks on manufacturing plants or production lines and not on targets that are not related to business. In other words, the type of target could make a difference on whether terrorism has an effect on FDI inflow. To prove their point, Powers and Choi (2012) distinguish between business-related attack and non-business-related attacks in their study. They find that, in their sample of 123 developing countries, transnational terrorist attacks on businesses have a negative impact on inward FDI stock, whereas terrorist attacks on non-business-related targets have no significant effect on FDI. Powers and Choi (2012) pooled all the other targets together into non-business-related terrorism. However, this could lead to a selectivity bias as attacks on transportation targets (e.g. railroads) or utility targets (e.g. power lines) could also disrupt business activities. Moreover, an attack on a government institution or on a police force could also make a location less desirable for foreign investors as it could decrease the perceived institutional quality or rule of law of a country. Hence, for this research I will expand on this thought that terrorism on different targets will have a different impact on the inward FDI.
2.2 FDI and the effects on the host country
7 These fluctuations in FDI are a concerning matter for countries that are dependent on their inward FDI flows. Host countries often stimulate to attract FDI as it is perceived as a vital part of a country’s economic development. From a micro perspective, the idea is that FDI provides access to advanced technology, skills, R&D and know-how to their countries. On a macro-level, FDI is believed to facilitate economic growth.
From the microeconomic perspective, it is expected that FDI affects the business environment in the host country as domestic firms will learn from the multinationals in the country due to spillovers.
Extending on the Melitz model (2003)2, Helpman, Melitz and Yeaple (2004) find that only the most
productive firms engage in FDI. Thus, increasing the aggregate productivity in that market. This phenomenon could lead to spillovers as domestic firms could learn from these productive MNCs entering the market in order to increase their productivity to survive in the market.
Blömstrom and Kokko (1998) distinguish between two types of spillovers: productivity spillovers and market access spillovers. The first spillover pertains to MNCs entry to a market leading to productivity and efficiency benefits for the domestic firms. The second spillover applies to MNCs assisting local firms to overcome the barriers to export markets by providing access to distribution networks, marketing outlets, and information about regulatory standards and consumer preferences (Blömstrom & Kokko, 1998). However, literature cannot come to a consensus on whether FDI from MNCs generates spillovers to
domestic firms.3
On a macro-level, much research has been done to investigate the impact of FDI on economic development and growth of a country. They find that FDI indirectly affects economic growth by positively impacting human capital (Li & Liu, 2005) and transfers of technology (Borenzstein, De Gregorio & Lee,
1998). Research on the direct effect of FDI on economic growth has shown some ambiguous results.4
According to Iamsiraroj (2016), these conflicting results have come from scholars ignoring the idea that FDI and economic growth have a dynamic relationship. In this relationship, FDI and economic growth simultaneously stimulate each other. In his research, Iamsiraroj (2016) finds that the overall effect of FDI is positively associated with growth and vice versa.
Although literature cannot come to a complete consensus on whether FDI actually benefits the micro and macro economy of a country, many states –especially developing countries- have become dependent on FDI inflows as a form of external finance (UNCTAD, 2018). Hence, a sudden decrease in FDI could still have detrimental effects on an economy. Therefore, it is important to understand how
2 The Melitz model suggests that exposure to trade prompts the most productive firms to enter the export market, whereas the least productive firms are forced to exist. Hence, increasing the aggregate productivity in an industry. This is due to the interaction between the firm heterogeneity in terms of productivity and the costs of exporting. Only the most productive firms are enabled to bear these fixed costs of exporting.
3 Görg and Greenaway (2004) provide a survey on the literature on spillovers in their paper. They find that there is no consensus on whether FDI triggers different type of spillovers. Therefore, they suggest that not every firm will experience the same effects from spillovers. Javorcik (2004) finds that positive spillovers only appear with projects that enjoy shared domestic and foreign ownership and not projects with fully foreign owned firms.
8 sudden events such as a terrorist attack could impact FDI and therefore directly or indirectly influence the economy of a country.
2.3 FDI and its determinants
To understand how terrorism could affect FDI inflows, it is necessary to first comprehend what drives a foreign firm or investor to choose a location over another. Governments around the world implement policies to enhance the ease of doing business to provide locational advantages. They have lowered entry barriers and opened up sectors to foreign investors in order to make the country more attractive for foreign investors (Brewer & Young, 1998).
There has been extensive research paired with additional policy advice for governments to detect which characteristics of a country are determinants of FDI. These determinants are often characterized as economic and institutional factors. Economic factors that are often studied to increase FDI are GDP growth, trade openness, exchange rate, and market size (Saini & Singhania, 2017). When considering the institutional factors, research often finds that countries with democratic institutions, political stability and rule of law attract FDI, whereas corruption, cultural distance and tax policies could deter FDI (Busse & Hefeker, 2007; Bailey, 2018).
However, there is no real consensus on which the “true” determinants are to attract FDI (Kok & Ersoy, 2009). A reason for this could be the difference in motivations for FDI. There are multiple motivations for a firm to engage in FDI. According to Dunning and Lundan (2008), companies engage in
FDI to pursue any of the following: markets, resources, efficiency and strategic assets.5 Helpman et al.
(2004) suggest that firms engage in FDI to access markets to avoid trade frictions. Ekholm, Forslid and Markusen (2007) argue that firms could use a host country as an export platform. These variations in motivation for FDI result in a disparity of determinants that a firm could pursue in a host country. For instance, a firm seeking markets could be more focused on a large market size when choosing a country, whereas a resource seeking firms could have a higher interest in countries with a relatively cheaper labor force.
In other words, firms are more prone to engage in FDI in countries that have locational advantages suiting the motivation for the FDI. Thus, some factors of FDI attractiveness can be more significant for some countries than for others. Recent studies on FDI determinants have made a distinction between FDI determinants for developed, emerging and developing countries, since the FDI determinants are likely to
differ between these groups of countries.6
Accordingly, Saini and Singhania (2017) find that in developed countries firms seek policy-related determinants as GDP growth, trade openness and economic freedom, whereas in developing countries
5 Market seeking firms want to exploit opportunities granted by foreign markets of greater dimensions. Resource seeking firms aim to acquire a type of resource that is either not available or more expensive in the home market. Efficiency seeking firms are companies that want to take advantage of heterogeneity in availability and costs of factor endowments, institutional arrangements, economic systems and policies, and market structures by concentrating production in a limited number of locations to supply multiple markets. Strategic asset seeking firms acquire assets of foreign corporations to promote their long-term strategic objectives (e.g. advancing their global competitiveness) (Dunning & Lundan, 2008).
9 MNCs look for economic determinants such as efficiency variables, trade openness, and gross fixed capital formation. Furthermore, Blonigen (2005) highlights that particularly in less-developed countries institutional quality is an important FDI determinant. Poor institutional quality required for well-functioning markets could increase the costs of doing business and lower the expected profitability. Hence, poor institutional quality would diminish FDI activity in a country.
Now that the literature on motivations and determinants of FDI has been discussed, the following section will continue on how terrorism could potentially affect the decisions of MNCs where to or where not to locate FDI. From this analysis, my hypotheses are constructed.
3. Theory Development
Dunning’s eclectic paradigm (1988), also called the OLI-framework, endeavors to explain why MNCs decide to internationalize through FDI. The OLI-framework refers to three requirements a firm needs to meet in order to engage in FDI: ownership advantages (O), locational advantages (L), and internalization (I). First, ownership advantage pertains to a firm needing to have a resource or advantage specific to the firm in order to offer extra value over their competition. Second, FDI will only be beneficial to the firm if there is some locational advantage in the host country that can be combined with the firm’s advantage. Third, there needs to be an internalization advantage which leads to the firm having more benefits in engaging in FDI rather than selling or leasing the ownership advantage.
Since ownership advantages are mainly dependent on the firm’s inner characteristics (Gastanaga, Nugent & Pashamova, 1998), it will be less affected by terrorism in a host country. However, terrorism could have a lasting effect on the latter two requirements. Terrorism in a host country could possibly take away the locational advantage, causing a firm to take their FDI elsewhere. Besides that, terrorism in a host country could make it less attractive to engage in FDI and more appealing to choose for other options such as exporting or licensing (Czinkota, Knight, Liesch & Steen, 2005); thus, taking away the internalization argument. Terrorism could lower the desirability of a location for investment in the following ways.
First, terrorism could decrease the market size of a country. Terrorist attacks can result into economic destruction in the sense that people will lose their jobs and trust. This could lead to decreasing incomes and consumption (Blomberg, Hess & Tan, 2011). Not to mention, terrorism could lead to more uncertainty in investments and savings (Crain & Crain, 2006). This decrease in the components of GDP could cause uncertainties for MNCs with regard to whether the location will be a good market for their products or that they will gain from their investment. Firms wanting to expand their markets through FDI could therefore choose a different location.
Second, terrorist attacks could diminish economic growth (Meierrieks and Gries, 2012). As discussed in the previous paragraph, GDP per capita can experience a downturn due to economic destruction. Additionally, the economic growth could decrease. Firms that invest to acquire new markets and to pursue growth opportunities abroad could therefore be less inclined to choose locations that lack or have a diminishing economic growth.
10 shipping delays. Even more, an attack on a power line or on an internet server could lead to a disruption in the production or distribution process, leading to shortages or disabling a firm of providing services. Firms that seek to serve a foreign market, to find efficiency or to use a market as an export platform all rely on good infrastructures. A country that has a higher perceived risk to have infrastructure disruptions due to terrorism will therefore become a less desirable location.
Fourth, transaction costs and the costs of doing business can increase due to terrorism. Terrorist attacks and the risk of terrorism could raise these costs in multiple ways. One of which is the security initiatives implemented by governments. Since 9/11, governments have undertaken more initiatives to enhance security in key systems such as logistics, communication, and infrastructure of international transportation (Czinkota et al. 2010). This increases costs for international firms; hence, these security efforts affect the competitiveness of these firms (Spich & Grosse, 2005). Moreover, labor costs might increase in terrorized areas in order to compensate the employees for the disamenities of working in a region dominated by terrorism (Frey et al. 2007). This rise in costs will diminish the efficiency or resource advantage a host country has. Furthermore, increasing costs could lead to firms choosing for cheaper options to access foreign markets such as exports.
Fifth, terrorism could decrease the trade openness of countries. Terrorism provokes governments to strengthen national security measures which increase the difficulty and costs of trade (Mirza & Verdier, 2014) as well as trade insecurity (Pham & Doucouliagos, 2017). Thus, countries that are hit by terrorism trade significantly less than countries not affected by terrorism (Nitsch & Schumacher, 2004). According to Helpman et al. (2004), trade frictions would lead to more firms engaging in FDI. However, companies that engage in FDI to use the host country as an export-platform rely on trade. Thus, these firms could be less inclined to invest in countries that experience terrorism to avoid trade frictions. The same arguments holds for firms that are resource-seeking and therefore utilize intra-firm trade to export the goods made in the host country to the home country. Hence, decreasing trade openness due to terrorism could result into a lower attractiveness of a location.
Last, terrorism increases the overall level of uncertainty and perceived risk in a country. This could lead to increased costs of investing in technology and capital formation (Blomberg et al., 2011). If the perceived risk increases, firms could decide to select safer locations. Furthermore, they will be more prone to choose for less riskier ways to infiltrate the foreign market such as exports or licenses.
In sum, terrorism could make a host country less attractive for a MNC. Hence, MNCs include terrorism in their decision-making regarding internalization and locating foreign operations (Li, Tallman & Ferreira, 2005). As illustrated by Ouyang & Rajan (2017), the frequency and intensity of terrorist attacks significantly diminishes merger & acquisition flows. Moreover, multiple studies found evidence that the number of terrorist attacks deter FDI stock (Enders et al., 2006; Abadie & Gardeazabal, 2008) and FDI inflows (Bandyopadhyah et al., 2014). Thus, I hypothesize the following:
H1: Terrorism in a host country has a negative effect on the FDI inflow of that country.
11 development levels will entail other results. Developing countries are often associated with a less-developed institutional frameworks. Therefore, they are anticipated to encounter more significant economic consequences when it experiences a terrorist attack than its better developed counterparts (Sandler & Enders, 2008; Meierrieks & Gries, 2012).
Better developed institutional frameworks will benefit countries managing terrorism in several ways. First, developed institutions will enable a country to prevent or monitor terrorism better. According to Li and Schaub (2004), underdevelopment is often linked to countries that provide safe havens for terrorists or that are unable to successfully expel terrorists from their border. Moreover, Bandyopadhyay and Younas (2011) found that better legal institutions will deter both domestic and transnational terrorist activities in a country. Furthermore, the war against terrorism is partly fought through the use of intelligence networks. Infinitively, we could assume that countries with developed institutions are capable to set up and manage more advanced intelligence networks which facilitates monitoring and preventing
terrorism.7 Second, governments in developed countries are better equipped to apply monetary, fiscal, and
other policies to recover from terrorist attacks (Sandler & Enders, 2008). For instance, an attack on a railroad that could cause delays might be faster rectified in a developed country than in a developing country.
This challenge experienced by developing countries to handle terrorist attacks could harm inward FDI in the following ways. First, it could increase the costs of doing business. The longer the infrastructure of a country is down due to a terrorist attack, the higher the costs will be to counter the disruptions in the supply chain. This will diminish the locational advantages concerning efficiency, which is an important FDI determinant for developing countries (Saini & Singhania, 2017). Firms that want to engage in efficiency-seeking FDI could now decide to locate their FDI elsewhere. Second, it takes longer to gain back trust from investors. According to Enders et al. (2006), developed countries are more capable to win back investors’ confidence through their security measures and economic policies. This implies that developing countries will encounter more difficulty to obtain the confidence back from investors after a terrorist attack has struck the country. Foreign investors, therefore, could decide to locate their FDI into a country in which they have more assurance to gain from their investments.
Moreover, institutional quality is an important factor for MNCs deciding to locate FDI in developing countries (Blonigen, 2005). Terrorism affects institutional quality by negatively affecting political and economic governance factors in less developed countries (Asongu & Nwachukwu, 2017). Thus, decreasing the attractiveness of the country for foreign investors.
In sum, developed countries often have better institutions to weather the economic consequences induced by terrorist attacks and to restore confidence. MNCs and other foreign investors could take this into consideration in their decision-making process on where to locate FDI. For instance, Ouyang and Rajan (2017) demonstrate in their paper that having a good institution -based on the Rule of Law indicator- significantly offsets the negative impact terrorism has on cross-border M&As. Hence, I hypothesize the following:
7 While reviewing literature on this topic, I stumbled upon an absence of research on the prevention and monitoring techniques
12
H2: Terrorism has a larger negative effect on the FDI inflow in a developing country than in a developed country.
It could be problematic to assume that all terrorist attacks will make a location less desirable for a MNCs decision to invest in that country. Instead, MNCs are prone to view certain types of terrorist attacks, by means of whom or what is attacked, as more detrimental to their business activities than others. Therefore, perceiving some locations as less desirable for investment when these countries encounter these more perilous attacks. This idea is aligned with Li’s suggestion that ‘less significant and limited terrorist attacks may have little effect on the expected return of an investment project’ (2006, p. 233).
Attacks on certain targets will likely influence inward FDI more than others. Intuitively, I assume these attacks to be more influential due to one of the following arguments. First, some targets could receive more media coverage (e.g. international tourists) which could increase the perceived risk in that country. Second, destruction of certain type of targets could cause more disruptions in the business activities of MNCs (e.g. attacks on transportation or utilities). Third, terrorist attacks on some targets could potentially diminish the institutional quality or rule of law in a country (e.g. attacks on government buildings, police force) which could make a country less attractive for investment if the MNC relies on the rule of law and institutions to do business. Fourth, attacks on some targets (e.g. diplomatic) could put constraints on the relationships between countries, which could make it more difficult for the MNC to do business or investments in a certain location.
Thus, terrorist attacks on certain targets could create more uncertainties and impose a bigger threat on MNCs’ objectives for their FDI. The only research that has taken this idea into consideration is Powers and Choi (2012). As I mentioned before in the literature review, they made a distinction between business-related and non-business-business-related terrorism. They find evidence that only business-business-related attacks have a significant impact on inward FDI. Thus, they are the first to provide some evidence on the idea that not all terrorist attacks will have an impact on inward FDI. Therefore, I hypothesize the following:
H3: The effect of terrorism on FDI depends on the type of target.
4. Methodology
4.1 Variables
Independent variable: Terrorism
Previous research on terrorism and its impact on FDI often utilize the number of transnational terrorist attacks as an independent variable. This probes two issues.
13 covers only transnational terrorist incidents, whereas GTD covers domestic and transnational terrorist incidents. However, ignoring domestic terrorism could lead to biases as the results would underestimate the actual impact of terrorism since domestic terrorist attacks outnumber transnational attacks (Enders, Sandler & Gaibulloev, 2011). Furthermore, GTD is considered to be more transparent compared to ITERATE in regards to the context and citation for sources of terrorism data (Sheehan, 2012). Some research has shown that the GTD has undercounted terrorist attacks prior to 1998 (Enders et al., 2011). However, since 2006 the National Consortium for the Study of Terrorism and Responses to Terrorism (START) took over the database and cleaned the data from 1998 to onward.
For this research, I use data of all terrorist attacks- domestic and transnational. Thus, I will be using the GTD. Given the previously discussed ambiguity of the data in the GTD dataset prior to 1998, I constrain the research with data from January 1998 till December 2017. Table 1 records the amount of terrorist attacks and casualties registered in the GTD from that period. This table shows that from half of
the attacks it is undecided or unknown to whether a terrorist attack is domestic or transnational.8 Given
this problem in the data, I have decided not to distinguish between domestic and transnational attacks within the main analysis. However, since previous research has mainly focused on transnational attacks, I will execute the regressions on the attacks considered to be transnational terrorism as a robustness check.
Table 1. Terrorist attacks and casualties registered in the GTD from 1998-2017
Domestic Transnational Unknown Total
Terrorist attacks 33502 20695 59987 114184
Casualties 229713 192106 259293 681112
Source: Global Terrorism Database (2018)
Second, using only the number of terrorist attacks could also cause for biases, since a large attack with many casualties will have a larger impact on the media and on the perception of the terrorism risk in that country (Frey et al. 2007). Hence, to provide a more complete picture on how terrorism affects FDI, it would be important to not only analyze the number of terrorist attacks, but also the number of casualties caused by terrorism. Therefore, I include both variables in my analysis: the number of terrorist attacks in a country and the number of casualties (fatalities and injuries) caused by terrorism.
Moreover, the GTD database provides more information on the nature of terrorist attacks. Thus, I also test specifically on how different targets of terrorist attacks could affect inward FDI. The GTD has specified 22 categories of targets.9 The ones used in the analysis are: private citizens and property,
military, police, government, business, transportation, religious figures and institution, educational institution, utilities, terrorist and non-state militias, journalists and media, government (diplomatic), violent political parties, NGO, telecommunication, airport and aircraft, tourists, food or water supply, and maritime.
8 The GTD database provides a dummy variable to indicate if an attack is transnational, domestic or unknown.
14 As the impact of terrorism is anticipated to be stronger for smaller than for larger countries (Sandler & Enders, 2008), I transform the indicators of terrorism to per 100000 inhabitants of that country. This will provide a better reflection of the perceived threat of terrorism in a country to a foreign investor (Bandyopadhyay et al., 2014). Similar studies have population-adjusted their terrorism indicators as well (Meierrieks & Gries, 2012; Bandyopadhyay et al., 2014).
Dependent variable: Inflow FDI flow
Foreign direct investments has two different measures: flows and stocks. According to the definitions provided by the OECD (2018), “inward flows represent transactions that increase the investment that foreign investors have in enterprises resident in the reporting economy less transactions that decrease the investment of foreign investors in resident enterprises”, whereas “inward FDI stock is the value of foreign investors' equity in and net loans to enterprises resident in the reporting economy”. Since I want to investigate whether terrorism in a country changes the attractiveness of a country to invest in, I will use the net inward FDI flows of a country provided by the United Nations Conference on Trade and Development (UNCTAD) database. Data on inward FDI flows contains skewness. In order to mitigate this skewness and reduce the impact of influential observations, I transform the inward FDI flows into natural log. However, some inward FDI flows contain negative data which will be regarded as missing data when transformed as a log. Thus, similar to Li (2006), I focus solely on the positive inward FDI
flows.10 To verify the robustness of my findings, I use logged inward FDI stock from the UNCTAD
database.
Moderating variable: Developing countries
To examine whether terrorism has a greater effect on FDI in developing countries, I use a dummy variable determining whether a country is a developing country or not. The value of the dummy variable equals 0 if it is a developed country, whereas it is a developing country when it equals 1. The development level of a country is based on the IMF classification from 2017 (International Monetary Fund, 2017). I analyze what effect the level of development has on the relationship between terrorism and FDI by using the development level of a country in an indicator variable.
Control variables
Besides the variables previously mentioned, I also utilize control variables to control for other factors that can impact the inward FDI of a country. I control for market size, economic growth, trade openness, institutional quality, infrastructure and lagged FDI.
A larger market size has shown to attract more FDI as it provides more opportunities to efficiently
use resources and to exploit economies of scale (Tsai, 1994; Asiedu, 2002). Based on previous research, a logged GDP per capita is taken to control for market size (e.g. Powers & Choi, 2012; Alam & Shah,
15 2013). Likewise, countries that experience economic growth have been associated with attracting more FDI. Foreign investors are drawn to countries that experience economic growth as it could lead to a higher level of aggregate demand and generate more profits from the realization of economies of scale (Iamsiraroj, 2015). To control for economic growth, the percentage change of GDP per capita is often used (Powers & Choi, 2012; Iamsiraroj, 2015).
Trade openness is considered to influence FDI. MNCs that go abroad to undertake market-seeking
investments could choose to engage in FDI instead of exports due to trade restrictions (thus, less trade openness); suggesting that trade openness would have a negative effect on FDI. However, MNCs that seek export-oriented investments could prefer to locate in economies that are more open to trade (Asiedu, 2002). Considering that multiple studies have found that trade openness positively influences FDI inflows, I hypothesize that trade openness of a country increases FDI inflows (Asiedu, 2002; Saini & Singhania, 2017). Trade openness is measured by the ratio of exports plus imports to GDP (Asiedu, 2002; Bandyopadhyay et al., 2014).
I also control for institutional quality. Multiple studies have shown that institutional quality has an
influence on FDI as good institutions decrease the risk and increase the ease of doing business in a country
(Bailey, 2018). I obtained data from multiple indicators11 and constructed one variable institutional quality
by performing a principal component analysis.12
Similarly, I control for infrastructure. A well-developed infrastructure increases productivity of
investments, market access and ease of doing business (Asiedu, 2002; Alam & Shah, 2013). As mentioned by Asiedu (2002), a good measure of infrastructure should account both availability and reliability. However, there is a lack of data on the latter requirement. Hence, I am constrained to use an indicator of infrastructure based on the availability. Multiple papers use the number of telephone subscriptions. By analyzing the data on this indicator, I found that telephone subscriptions are declining and becoming less relevant in the last few years of my sample. Thus, I use the percentage of internet users in a country instead. Access to internet remains a relevant way to obtain access to goods and services. Differences in other infrastructure indicators such as access to a port will be accounted for by including country fixed effects.
Lastly, I include a lagged variable of inward FDI flow (and FDI stock in the robustness check) this is in line with other studies examining the terrorism-FDI link (Powers and Choi, 2012; Bandyopadhyay et al. 2014). The lagged FDI variable is an important factor to correct for autocorrelation. It should have a positive impact on FDI inflows (Saini & Singhania, 2017).
4.2 Sample
The sample for this analysis is based on data available in the GTD from 1998 until 2017. From this time period the GTD has data from terrorist events occurring in 168 countries. With comparing the data from these countries with data retrieved from the UNCTAD and World Bank for the dependent and control variables, I excluded 13 countries which showcased problematic issues regarding their lack of data. These
11 These indicators are control of corruption, government effectiveness, rule of law, regulatory quality, political stability, and voice and accountability taken from The Worldwide Governance Indicators (WGI) project from the World Bank.
16 included countries that do are not included in one of the three databases (e.g. Taiwan and Cuba), countries that barely have had any data on them in the last 20 years (e.g. South Sudan, Swaziland) and countries that have experienced a separation of states in the last 20 years (e.g. Yugoslavia, Serbia-Montenegro). I ended up with 155 countries in my analysis over 20 years. The list of countries is presented in Table 6 in Appendix A.
4.3 Model specifications
In this section, I will discuss the models I use to examine my hypotheses. A representation of the variables and their sources are below in Table 2.
Table 2. Variables, definition and source
Variable name Variable code Definition Source
Terrorist attacks TA Frequency (number) of terrorist attacks per 100000 inhabitants
GTD
Casualties TC Intensity of terrorism: casualties(fatalities and injured) per 100000 inhabitants
GTD
Terrorist attacks per target type
Type TA Number of terrorist attacks on a certain target type per 100000 inhabitants. Target types: Private, Tourism, Business, Transportation, Religion, Education, Utilities, Terrorist, Media, Diplomatic, Government, Violent Politics, NGO, Telecommunication, Air, Police, Maritime, Food and water supply, and Military
GTD
FDI stock FDIstock Log of inward FDI stock UNCTAD
FDI flow FDIflow Log of inward FDI flow UNCTAD
Market size GDP Log of GDP per capita of host country in US dollars World Bank Economic growth EG Annual percentage growth rate of GDP World Bank Trade openness Trade Volume of trade as a percentage of GDP World Bank Institution quality Institutions Principal Factor Component of the following
governance factors: Voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, control of corruption
World Bank
Infrastructure Internet Individuals using the internet (% of population) World Bank Development level Developing Dummy variable taking on the value 1 if country is a
developing country
IMF
Before conducting the regressions, I have transformed GDP, FDI stock and FDI flow into natural logs in
order to mitigate the influence of high differences between observations.13 Furthermore, I have
interpolated missing values and checked for outliers. A descriptive summary and correlation matrix are presented in Table 7 and 8 in Appendix B. To analyze the previous suggested hypotheses, I conduct within
13 No logs were taken from the variables casualties and number of terrorist attacks since the data contains a large amount of
17
estimations consisting of OLS panel regressions with fixed effects.14 In order to counter heteroscedasticity
in the data, I use robust standard errors in the regression.
I will start the analysis with a model that examines the effect of the aforementioned control variables on the FDI inflow:
(1) ln(FDIflow it) = α + β1 (ln(GDPit)) + β2 (EGit) + β3 (Tradeit) + β4(Institutionit) + β5 (Internetit) + β6
(FDIflowit-1) + γt + γi + εit
For the control variables, I expect market size (GDP), economic growth (EG), trade, institutional quality (Institution), of country i at time t to all have a positive effect on the FDI inflows. Furthermore, I include
a lagged dependent variable (FDIflowit-1) to control for the autocorrelation within the data. I expect the
lagged FDI flow to be positive as well. In every model, I control for country (γi) effects to control for
factors that do not change over time for any given country. Likewise, I control for time (γt ) fixed effects
to capture the influence of aggregate time-series trends on inward FDI. Fixed effects will, thus, absorb the influence of unobservable factors on FDI.15
Next, I will examine hypothesis 1; how terrorism affects FDI inflows. Model 2 examines whether the number of terrorist attacks (TA) in country i at time t influences the FDI inflow in that country in the same year:
(2) ln(FDIflow it) = α + β1 (TAit) + β2 (ln(GDPit)) + β3 (EGit) + β4 (Tradeit) + β5 (Institutionit) + β6 (Internetit) + β7 (FDIflowit-1) + γt + γi + εit
As I expect terrorism to have a negative effect on inward FDI, I anticipate the coefficient of TA to be negative (β1<0). The same model is used to examine whether the number of casualties induced by terrorism (TC) influences FDI inflows. TC will then replace TA in the model. Similarly, I expect a negative coefficient for TC (β1<0).16
Model 3 is utilized to examine whether the development level of a country has a moderating effect on the relationship described in model 2. In other words, whether the effect of terrorism on FDI in a developed country is different than in a developing country. This will be examined by including the dummy variable Developing which indicates whether a country is developed (base-value) or developing (value equals 1). Furthermore, an interaction term is included. This leads to the following model:
14 To be sure I conducted a preliminary Hausman test. This test shows that there is heteroskedasticity; thus, the estimated coefficients are different. Fixed effects is then preferred over random-effects models as the latter could generate inconsistent estimates. To check for the robustness of my findings, I also ran the regressions with random effects. These tests did not lead to different conclusions. These tests are presented in the log file.
15 FDI restrictiveness has been considered as a control variable, but there is a lack of data on this variable concerning the sample. However, country fixed effects will control for the unchanged FDI restrictiveness in countries over time.
18 (3) ln(FDIflow it) = α + β1 (TAit) + β2 (Developingi) + β3 (Developingi*TAit) + β4 (ln(GDPit)) + β5 (EGit)
+ β6 (Tradeit) + β7 (Institutionit) + β8 (Internetit) + β9 (FDIflowit-1) + γt + γi + εit
As hypothesized, I expect that terrorism has a larger negative effect on developing countries. Thus,
I anticipate that the coefficient of TA to be negative (β1<0) and the coefficient of the interaction variable
to be negative (β3<0). Similarly to model 2, I will examine the same model for TC and I expect a similar
outcome.
To further analyze how terrorism influences FDI inflows in developed and developing countries, I will examine the nature of terrorism regarding the target type of the terrorist attack. More specifically, I will examine whether different types of targets have different effects on FDI inflow. I will separate this in three types of models: one for worldwide, one for developed countries and one for developing countries. This generates the following models:
(4) ln(FDIflow it) = α + β1 (Type TAit) + β2 (ln(GDPit)) + β3 (EGit) + β4 (Tradeit) + β5 (Institutionit) +
β6 (Internetit) + β7 (FDIflowit-1) + γt + γi + εit
(5) ln(FDIflow it) = α + β1 (Type TAit) + β2 (ln(GDPit)) + β3 (EGit) + β4 (Tradeit) + β5 (Institutionit) +
β6 (Internetit) + β7 (FDIflowit-1) + β8 (Developing=0i) + γt + γi + εit
(6) ln(FDIflow it) = α + β1 (Type TAit) + β2 (ln(GDPit)) + β3 (EGit) + β4 (Tradeit) + β5 (Institutionit) +
β6 (Internetit) + β7 (FDIflowit-1) + β8 (Developing=1i)+ γt + γi + εit
Since the three target types with the highest frequency (private citizens and property, police, and military) have a high correlation with some of the other target types, I have taken them separate into three distinct regressions. A correlation matrix regarding the target types is presented in Table 9 in Appendix B. The other target types will be combined in one regression. I expect the target types to obtain different coefficients from one another.
5. Results
5.1 Basic results
Terrorism and development level
Table 3 and 4 report the models discussed in the previous section of the paper. Table 3 includes the models used to examine the first two hypothesis. It shows that both the number of terrorist attacks and the number of casualties have a negative, albeit insignificant effect on inward FDI flows.
19 effect on FDI inflows in developed countries and that this negative effect diminishes when a country is less developed.
Table 3. The effect of terrorism on inward FDI flow (log), 1998 – 2017. Fixed-effects regression
Worldwide Development level
Variable Model 1 Model 2 Model 3 Model 4 Model 5
Number of terrorist attacks 0.016 -0.160
(0.077) (0.097)
Number of casualties -0.007 -0.028***
(0.006) (0.007)
Developing*number of terrorist attacks 0.190
(0.130)
Developing*number of casualties 0.024**
(0.010)
Developing Omitted Omitted
Market size (log) 0.494*** 0.493*** 0.497*** 0.489*** 0.499*** (0.099) (0.099) (0.100) (0.100) (0.099) Economic growth 0.014*** 0.014*** 0.014*** 0.014*** 0.014*** (0.005) (0.005) (0.005) (0.005) (0.005) Trade openness 0.003*** 0.003*** 0.003*** 0.003*** 0.003*** (0.001) (0.001) (0.001) (0.001) (0.001) Infrastructure -0.013*** -0.013*** -0.013*** -0.013*** -0.013*** (0.002) (0.002) (0.002) (0.002) (0.002) Institution Quality 0.243*** 0.244*** 0.240*** 0.245*** 0.240*** (0.056) (0.055) (0.056) (0.055) (0.056) Lagged FDI flow 0.364*** 0.364*** 0.363*** 0.364*** 0.364***
(0.032) (0.032) (0.032) (0.032) (0.032)
Constant -0.117 -0.110 -0.127 -0.083 -0.147
(0.714) (0.718) (0.719) (0.721) (0.718)
Country effects Yes Yes Yes Yes Yes
Time effects Yes Yes Yes Yes Yes
N 2659 2659 2659 2659 2659
adj. R-square 0.519 0.519 0.519 0.519 0.519
BIC 5898.932 5906.725 5906.054 5913.671 5913.003
rss 1332.923 1332.877 1332.541 1332.406 1332.072
Note: Robust standard errors in parentheses in the second row; ***, **, * denote significance at 1%, 5% and 10%
respectively; Adjusted R-square is taken from the within R-square; Number of terrorist attacks and casualties is per 100000 inhabitants of that country.
This table shows three results. First, the number of terrorist attacks has no effect on inward FDI flow. Second, the number of casualties only has the predicted negative effect on developed countries
and not on developing countries.Third, to magnify the last point, the interaction variable demonstrates
20
Regarding the control variables, market size, economic growth, trade openness, institution quality,
and lagged FDI flow, all show the expected positive and significant effect on inward FDI flow in Table 3. Infrastructure, however, demonstrates a significant negative effect on inward FDI flow which is opposite of what is anticipated. This result could be due to internet not being the correct measure for infrastructure given that it is a part of quantity and not part of the quality of infrastructure (Asiedu, 2002).
Type of targets
Table 4 demonstrates the results regarding the effect of terrorist attacks on different types of targets on FDI inflow. Models 1 to 4 show the results for all the countries combined, in which government-related, and tourism-related targets have a significant and negative effect on FDI inflow, whereas an attack on an education-related target has a significant, positive effect on FDI inflow. Models 5 to 8 conclude that in developed countries terrorist attacks on transportation-related, diplomatic-related and maritime-related targets negatively affect FDI inflows, while attacks on targets regarding the police, violent political parties and aviation positively influence FDI inflow. Lastly, models 9 to 12 pertain to developing countries. These models record that attacks on military-, government- and tourist-related targets have a negative effect on FDI inflows. Attacks on targets concerning maritime or educational institutions, on the other hand, positively impact FDI inflow.
Nonetheless, most of the included targets do not show any significant relationship with inward
FDI flow. In other words, terrorist attacks on these targets do not influence the desirability of a location for foreign investors. This is in line with the argument of the paper that not all terrorist attacks will
culminate into a decrease in inward FDI.17
Furthermore, Table 4 presents evidence that the results of the analysis on target types differ
between developing and developed countries. As both military and government-related attacks illustrate significant negative coefficients for developing countries, it could be suggested that MNCs or foreign investors will be more reluctant in developing countries if their institutional quality and rule of law is impaired.
In short, Table 4 provides support for hypothesis 3: terrorist attacks on different targets have
different effects on inward FDI. In other words, not every terrorist attack will affect MNCs decisions to locate their investments elsewhere and therefore decrease inward FDI flows in that country. Moreover, it shows that FDI flows in developing and developed countries are affected by different terrorist attacks.
17 Extraordinary is that business-related targets do not show any negative significant relationship as it intuitively would be
21
Table 4. The effect of different type of targets of terrorism on inward FDI flow (log), 1998-2017.
Fixed-effects regression
Worldwide Developed countries Developing countries
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Private 0.118 -0.141 0.205 (0.235) (0.218) (0.260) Police 0.298 5.099* 0.228 (0.294) (2.636) (0.263) Military -0.294* 0.190 -0.349** (0.161) (0.483) (0.159) Government -1.000** -0.276 -1.286*** (0.467) (1.993) (0.485) Tourism -2.015* -23.927 -2.217** (1.028) (19.984) (1.055) Transportation -0.680 -7.797*** 0.972 (1.797) (2.600) (1.611) Utilities -2.099 -12.226 -3.075 (1.998) (8.553) (1.891) Business 1.205 -0.362 2.082 (1.283) (1.686) (2.015) Diplomatic -1.195 -5.785* -0.977 (1.375) (2.886) (1.495) Terrorist -0.062 18.868 -0.177 (0.701) (14.933) (0.650) Telecommunication 3.278 6.925 2.530 (4.597) (17.366) (4.395) Violent Political Parties 1.230 19.776** 0.582 (2.190) (8.609) (2.431) Religion -1.936 -0.806 -2.517 (1.796) (1.025) (3.108) Food or water supply -22.214 -19.838 -31.396 (15.841) (13.520) (22.639) Media -0.010 -0.531 -0.347 (0.565) (1.329) (0.535) Maritime 4.992 -134.314* 23.786** (17.531) (78.126) (11.418) NGO -0.216 28.997 0.002 (5.074) (17.972) (4.946) Aviation -3.158 31.770** -2.435 (4.706) (12.636) (4.839) Education 5.468*** 11.751 4.376** (1.624) (8.849) (2.112)
Market size (log) 0.493*** 0.490*** 0.497*** 0.503*** 0.771*** 0.763*** 0.780*** 0.858*** 0.340*** 0.341*** 0.346*** 0.360*** (0.098) (0.098) (0.100) (0.102) (0.248) (0.252) (0.247) (0.262) (0.106) (0.105) (0.107) (0.108) Economic growth 0.014*** 0.015*** 0.014*** 0.013** 0.021 0.019 0.021 0.019 0.012** 0.012** 0.011** 0.011** (0.005) (0.005) (0.005) (0.005) (0.014) (0.014) (0.014) (0.014) (0.005) (0.005) (0.005) (0.005) Trade openness 0.003*** 0.003*** 0.003*** 0.003*** 0.002 0.002 0.002 0.003 0.003*** 0.003*** 0.003*** 0.004*** (0.001) (0.001) (0.001) (0.001) (0.004) (0.004) (0.004) (0.003) (0.001) (0.001) (0.001) (0.001) Infrastructure -0.013*** -0.013*** -0.013*** -0.013*** -0.009 -0.010 -0.009 -0.013* -0.009*** -0.009*** -0.010*** --0.010*** (0.002) (0.002) (0.002) (0.002) (0.007) (0.007) (0.007) (0.007) (0.002) (0.002) (0.002) (0.002) Institution Quality 0.245*** 0.247*** 0.238*** 0.234*** -0.091 -0.052 -0.094 -0.099 0.302*** 0.301*** 0.292*** 0.293*** (0.055) (0.055) (0.056) (0.058) (0.173) (0.156) (0.173) (0.141) (0.056) (0.056) (0.057) (0.060)
Lagged FDI flow 0.364*** 0.363*** 0.364*** 0.357*** 0.175** 0.168** 0.174** 0.160** 0.384*** 0.383*** 0.382*** 0.375*** (0.032) (0.032) (0.032) (0.033) (0.069) (0.070) (0.069) (0.071) (0.038) (0.038) (0.038) (0.039)
Constant -0.109 -0.087 -0.132 -0.140 -0.165 -0.125 -0.245 -0.832 0.820 0.818 0.793 0.719 (0.713) (0.715) (0.722) (0.736) (2.093) (2.136) (2.084) (2.265) (0.734) (0.734) (0.742) (0.746)
Country effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Time effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 2659 2659 2659 2659 606 606 606 606 2053 2053 2053 2053
adj. R-square 0.519 0.520 0.520 0.521 0.198 0.203 0.198 0.201 0.601 0.602 0.602 0.604
BIC 5906.308 5903.830 5904.944 6001.094 1567.244 1563.513 1567.363 1612.759 4361.335 4360.738 4359.798 4448.030
rss 1332.668 1331.427 1331.985 1320.942 361.732 359.512 361.803 350.825 916.603 916.337 915.918 904.315
Note: Robust standard errors in parentheses in the second row; ***, **, * denote significance at 1%, 5% and 10% respectively; Adjusted
22
5.2 Robustness check
In order to examine whether the previous results are robust, I have conducted three extra tests. First, I have substituted inward FDI flow for inward FDI stock. Second, I have run the regressions on transnational terrorism. Third, I have examined the lagged effects of terrorist attacks on FDI.
Inward FDI stock
The results of this analysis are presented in Table 10 and 11 in Appendix C. Table 10 shows how terrorism affects inward FDI stock. In particular, it demonstrates that the number of terrorist attacks has a negative, albeit insignificant impact on FDI stock even when the development level is taken into consideration. Nonetheless, the number of casualties does have a negative and significant effect on inward FDI stock. Furthermore, the development indicator variable shows that this negative effect is higher in developed countries than in developing countries. These results are in line with previous recorded results regarding FDI flows. Similarly, Table 11 provides evidence that attacks on certain type of targets impact inward FDI differently than attacks on other target attacks. Thus, the analysis on FDI stock also provides support for hypothesis 3.
Nonetheless, there are some different results regarding which target types impacts inward FDI. As shown in Table 11, terrorist attacks on business, government, transportation and diplomacy targets have a significant negative effect on FDI stock in developed countries. In developing countries, terrorist attacks on tourists or tourist attractions entail a significant and negative impact on inward FDI stock. These results show some discrepancies with the previously recorded results on inward FDI flow. Nevertheless, this is not extraordinary as stock and flows are measured differently. Stock is the equity of foreign investors measured at a point in time, whereas flows are the transactions of foreign investors measured over an interval of time. Thus, a decrease in stocks would refer to a capital investment being destroyed or taken away by the investors, whereas a decrease in flows could indicate that foreign investors are directing their investments to other more attractive locations. Furthermore, in my analysis I have 222 more observations in the regressions regarding FDI stock than in those for FDI flow. In order to transform the variables into logs, I had to delete the observations that were zero or negative. This led to a higher reduction of observations concerning FDI flows.
Transnational terrorism
Thus far, I have combined transnational and domestic terrorism into one variable of terrorism. Since previous research mostly focused on solely transnational terrorism, I have run the regressions of hypothesis 1 and 2 on only transnational terrorism data as well. These results are illustrated in Table 12 in Appendix D. Transnational terrorist attacks as well as casualties induced by transnational terrorism both have a negative and significant effect when the indicator variable is taken into the equation. This shows that both indicators of transnational terrorism have the predicted negative effect in the baseline countries, which are the developed countries.
23 Table 13 in Appendix D shows how transnational terrorist attacks on different target types impacts inward FDI flows. Similar to Table 4, Table 13 provides support for hypothesis 3 that some targets have an effect on inward FDI, whereas others do not have a significant impact. In comparison with the results from all terrorist attacks presented in Table 4, transnational terrorism elicits more targets to have an effect on inward FDI for developing countries and therefore also worldwide. The most extraordinary differences are the following. First, attacks on the government in developing countries and on diplomatic targets in developed counties both loose its significance with transnational terrorism. Second, religious and food or water supply targets obtain a significant negative coefficient when merely including transnational terrorist attacks. Third, the coefficient of educational institutions worldwide transformed from a significant, albeit positive into a significant negative coefficient. These results suggests that there could be a difference between transnational and domestic terrorism and the way it influences foreign investors’ FDI decisions.
Lagged effects
Table 3, only examines the immediate, short-term, effect of terrorism on inward FDI flow. However, if a country has a terrorist attack, it could take some time for MNCs to locate their FDI elsewhere. Thus, I also examined the effects of lags of 1, 2 and 3 years. These results are presented in Appendix E. Table 14 presents the lagged effects of terrorist attacks. I find a lagged terrorist attack of two years has a negative and significant impact on FDI flow worldwide. Thus, suggesting that worldwide it takes two years for foreign investors or MNCs to let a terrorist attack to influence their decision concerning FDI. More noteworthy, are the results in Table 14 that show that a lagged number of terrorist attacks of three years has a positive and significant effect on worldwide, developing, and developed countries. This suggests that the higher the number of a terrorist attack in a country, the higher the increase in inward FDI flows of that country after three years. This could indicate that when countries are able to reduce the amount of terrorism in their countries, investors and MNCs will gain back their confidence and start investing in the country.
This line of thought cannot be implied for the number of casualties as Table 15 shows. Table 15
24
6. Discussion
6.1 Discussion of results
This research provides evidence that terrorism does affect inward FDI in certain ways. The number of casualties induced by terrorism has a negative effect on inward FDI in developed countries. This effect, however, mitigates when developing countries are taken into the equation. There is no evidence that the number of terrorist attacks impacts FDI. However, when we distinguish between the targets of attacks, some terrorist attacks on certain targets do influence inward FDI flows negatively. Thus, I can conclude that either depending on the intensity of the terrorist attack and development level of a host country or on the type of attack target, terrorism could negatively impact inward FDI flows.
In the first analysis, there is no evidence that the number of terrorist attacks or the number of casualties worldwide or for developing countries has a negative impact on inward FDI. Only after including lagged variables of casualties of three years, the number of casualties shows a small significant negative effect on FDI inflow worldwide and in developing countries (see Table 15). This could imply that only when the intensity of terrorism has been increasing in previous years, foreign investors become more reluctant to invest in a developing country.
Not being able to find any significant results concerning the number of terrorist attacks could indicate that not every terrorist attack will influence foreign investors’ location decisions regarding FDI. This argument is corroborated with the results pertaining to the third hypothesis which investigates whether target types matters in terms of terrorism affecting inward FDI. Terrorist attacks on certain targets will influence the decisions of foreign investors as it will likely increase the perceived risks of investing in that country. For instance, an attack on a government building in a developing country could increase the political instability and therefore decrease the ease of doing business for MNCs. An attack on a newspaper, on the other hand, will less likely interfere with the objectives of foreign investors.
However, the results of Table 4 concerning target types also unexpectedly showed that attacks on certain target types have positive effects on FDI inflow. This would suggest that an attack on a police-related target, violent political parties-police-related or aviation-police-related target in a developed country would increase inward FDI flows. Similarly, an attack on targets related to maritime and education show a positive effect on inward FDI in developing countries. Attacks on aviation and maritime might could boost the use of FDI in order to avoid the increasing costs of exports. This would support the argument of Helpman et al. (2004) that trade frictions could lead to more FDI. However, these results could also suggest that foreign investors and MNCs could increase investments in countries experiencing terrorism in order to assist with the recovery or to counter the threats of terrorism. For instance, after an attack on a police force, a company could decide to invest into an extra subsidiary or take over a company that is concerned with developing a new security system in order to secure and maintain their business activities in that country (Jain & Grosse, 2009).
Lastly, I want to address the unanticipated results regarding hypothesis 2. I expected terrorism to
