Sage Journals: Discover world-class research

Abstract

This article presents an eclectic review of the analytical study of terrorism that views all agents as rational decisionmakers. This analytical literature began in earnest with the seminal study of US skyjackings by William Landes in 1978. After 11 September 2001, the analytical literature on terrorism grew rapidly. Based on policy relevance, my survey article identifies five key areas of intense research interests. These include analyses of terrorist attack trends, the economic consequences of terrorism, the study of counterterrorism effectiveness, the causes of terrorism, and the relationship of terrorism and liberal democracies. New developments in the field focused on distinguishing key differences between domestic and transnational terrorism. Additionally, recent game-theoretic advances permitted more active agents and stages to the games. Other major developments involved the study of networked terrorists and the role of counterterrorism foreign aid. Fruitful future directions include using advanced econometric methods to discern the true impact of terrorism on growth, applying spatial econometrics to the study of terrorism, ascertaining the determinants of terrorist groups’ longevity, and learning how to foster international counterterrorism cooperation.

Keywords

counterterrorism effectiveness domestic and transnational terrorism economic consequences of terrorism game theory

Introduction

Terrorism is the premeditated use or threat to use violence by individuals or subnational groups to obtain a political or social objective through the intimidation of a large audience beyond that of the immediate victims (Enders & Sandler, 2012). This definition is consistent with others in the literature (Hoffman, 2006; RAND, 2012). Violence is a hallmark of terrorism, with some terrorist groups engaging in gruesome attacks to create widespread anxiety or revulsion. To qualify as terrorism, an attack must have a political motive. By limiting terrorism to subnational agents, the above definition rules out state terror, where a government terrorizes its own people. However, the definition does not eliminate state-sponsored terrorism, in which a government secretly aids a terrorist group through funds, intelligence, safe passage, or some other means. Finally, the definition emphasizes that the true target of the anxiety-generating attacks is a wider public, who may pressure the government to concede to the demands of the terrorists.

The study of terrorism has been an active field of research in economics, political science, sociology, and related disciplines since the late 1960s. At first, political science took a conceptual and historical approach to the study of terrorism, which was pragmatic in the absence of data and theoretical constructs for terrorist behavior. The early conceptual approach focused on the definition of terrorism, the myriad causes of terrorism, the tactics of the terrorists, and the identity of the primary terrorist groups and movements (e.g. Crenshaw, 1981; Wilkinson, 1986).

The analytical study of terrorism derived, in part, from scholars, such as William Landes, who viewed terrorists as rational actors. Rationality is based on predictability rather than on the desirability of terrorists’ tactics or goals. If changes in terrorists’ constraints, say through government policies, result in predictable behavioral responses by the terrorists, then terrorists are rational actors. Some researchers – most notably Abrahms (2008) – questioned whether terrorists are rational, because they usually do not achieve their sought-after goal. Abrahms (2008: 83) stated: ‘I found that in a sample of twenty-eight well-known terrorist campaigns, the terrorist organizations accomplished their stated policy goals zero percent of the time by attacking civilians’. In contrast, Jones & Libicki (2008) indicated 132 campaigns where the terrorist groups renounced terrorism and joined the political process or else achieved their goal. Although Abrahms (2008) raised many interesting challenges to rationality, his evidence is somewhat selective.¹

The analytical study of terrorism was also fostered by the collection of event data. In his landmark study of skyjackings, Landes (1978) used US Federal Aviation Association (FAA) data on skyjackings to estimate the deterrent effects of US antiterrorism policies against skyjackings during 1961–79. Landes (1978) applied methods and concepts from the economic analysis of crime to identify effective deterrents to US skyjackings – e.g. the likelihood of apprehension, the likelihood of conviction, length of prison sentences, the presence of airport metal detectors, and sky marshals on planes. Most important, he found that airport metal detectors reduced US skyjackings by 4.4 to 8.8 incidents per year. Landes could then translate these findings into the value of metal detector installation.

After the heinous attacks on 11 September 2001 (9/11), there was an explosion of terrorism literature, both conceptual and analytical. The main purpose of this article is to present an eclectic overview of the analytical approach to the study of terrorism. The analytic literature is voluminous, so I must selectively choose which articles to cite. I favor articles that have either garnered a lot of citations or else, I believe, will do so in the future. This survey is organized as follows. In the next section, three primary terrorist event datasets are briefly introduced, because most of the empirical literature used one or more of them. In the ensuing five sections, the article highlights active areas of research that have much policy relevance. First, a knowledge of terrorist trends can inform forecasting and policy. Second, analyses of the economic consequences of terrorism identify some of the social costs of terrorist campaigns. Knowledge of these costs can indicate the potential gains from effective counterterrorism policies that limit terrorist attacks. Third, governments must ascertain the effectiveness of their defensive and proactive countermeasures, so that ineffective measures are discontinued. Fourth, understanding the root causes of terrorism allows a society to take steps to reduce grievances and, therefore, limit terrorism. Fifth, the relationship between democracies and terrorism can educate policymakers on the inherent risks of political regimes at home and in countries where these policymakers have vital interests. The next-to-last section presents some major new developments in the analytical study of terrorism – e.g. deconstructing terrorism into domestic and transnational attacks, game-theoretic analyses of counterterrorism, and the role of foreign aid. These developments are highlighted because there are lots of recent articles on these topics. Moreover, I believe that interest in these topics will continue in the future. The final section contains future directions for research and a few concluding remarks.

Event databases

An important distinction for the three event databases concerns domestic and transnational terrorism. Domestic terrorist incidents are homegrown and home directed with consequences for just the host or venue country, its institutions, citizens, property, and policies.² For a domestic terrorist incident, the three key agents – the perpetrators, victims, and targets – are all from the venue country. In contrast, transnational terrorist attacks involve victims, perpetrators, targets, supporters, or territory from more than one country. As such, transnational terrorism generates international externalities that confound effective policy.

The empirical research on terrorism has profited greatly from event databases that record key variables of terrorist incidents. Since the 1970s, the International Terrorism: Attributes of Terrorist Events (ITERATE) dataset has coded many variables (e.g. incident date, country location, target entity, attack type, casualties, perpetrators’ nationalities, terrorist group, victims’ nationalities, and logistical outcome) for transnational terrorist incidents. Currently, ITERATE’s coverage is 1968–2011, with yearly updates coming during the summer (Mickolus et al., 2012). ITERATE, like the other event databases, relies on the news media for its variables. For hostage-taking incidents, ITERATE also contains negotiation variables, which have been invaluable in studies on hostage taking (e.g. Gaibulloev & Sandler, 2009). The initial fixation of empirical articles on transnational terrorist incidents occurred because ITERATE was the most extensive dataset available during the 1980s and 1990s. Another competing dataset is the RAND (2012) terrorist event database, which currently codes terrorist incidents for 1968–2009. After 1997, the RAND dataset began recording domestic terrorist incidents, so that it identifies transnational terrorist incidents for 1968–2009 and domestic terrorist incidents for 1998–2009. Compared to ITERATE, the RAND event dataset codes fewer variables and for transnational terrorist incidents has more limited coverage than ITERATE (Enders, 2007).

The Global Terrorism Database (GTD) records both domestic and transnational terrorist incidents (La Free & Dugan, 2007), which could assist researchers in identifying differences between the two types of terrorism – e.g. their impacts on economic growth. For GTD, this partition of domestic and transnational terrorist incidents was first accomplished by Enders, Sandler & Gaibulloev (2011) for 1970–2007 and has been updated by them through 2010. There are many essential differences between GTD and ITERATE beyond the coverage of domestic terrorism by GTD. For example, the perpetrators’ nationalities are not recorded in GTD after 1997; very few logistical failures are listed in GTD; and many insurgent attacks are included before 1998 in GTD. These datasets can serve many useful purposes provided that researchers are aware of their idiosyncrasies.

How has terrorism changed since 1968?

Because event datasets only go back to 1968, any study of the patterns of terrorism must commence with 1968. Terrorism data show specific cycles (i.e. regular-spaced peaks and troughs) attached to alternative modes of attacks as terrorists and their government adversaries respond to one another’s countermeasures. As a general rule of thumb, more complex attack modes – e.g. skyjackings – display longer cycles than simpler modes – e.g. threats and hoaxes or bombings (Sandler & Enders, 2004). Insofar as bombings are a composite of many different kinds of attacks (e.g. letter bombs, explosive bombings, and car bombings), bombings are represented by multiple cycles of alternative lengths. By knowing these cycles, law enforcement agents may be able to anticipate the next wave of attacks when cycles are regular.

Transnational terrorist attacks rose from 1968 to the mid-1980s when there were about 500 incidents per year, as shown in Figure 1 based on ITERATE data. The 1980s was the decade of state-sponsored terrorism (Hoffman, 2006). After the mid-1990s, the number of transnational terrorist incidents fell precipitously to approximately 100 to 200 events per year. This fall is attributed to the decline in state sponsorship and the decline of left-wing terrorists. In Figure 2, I display the proportion of transnational and domestic terrorist events with casualties (i.e. deaths and/or injuries), based on ITERATE and GTD, respectively.³ As shown, transnational terrorism shows a rising proportion of incidents with casualties from 1997 on, corresponding to the growing dominance of the religious fundamentalist terrorists, who sought greater carnage (Enders & Sandler, 2000; Rapoport, 2004). Before 1990, 26% of transnational terrorist attacks ended in casualties; after 1990, 41% of these attacks resulted in casualties. Based on Enders, Sandler & Gaibulloev’s (2011) division of GTD data, Figure 2 also indicates that the proportion of domestic terrorist events with casualties displays a rising trend from 1989 to 2005. As shown, the proportion of domestic terrorist attacks with casualties is always greater than that of transnational terrorist attacks with casualties. Moreover, domestic terrorist incidents outnumber transnational terrorist incidents by four to ten times depending on the dataset and time period consulted.

Another interesting pattern comes from the terrorists’ choice of targets. Suppose that targets are partitioned into four general categories: officials, military, business, and private parties. As shown in Figure 3 (based on ITERATE), transnational terrorists have rationally responded to security measures in their targeting decisions. As officials were afforded protection by the state, transnational terrorists increasingly attacked softer business targets in the 1970s. As businesses protected their assets, terrorists switched to attacking private parties, so that these

Figure 1.

Number of transnational terrorist incidents (ITERATE), 1968–2011

Figure 2.

Proportion of transnational and domestic terrorist events with casualties

latter attacks surpassed business attacks in 1980 (Brandt & Sandler, 2010, 2012). By 1999, terrorist attacks against private parties even surpassed those directed at officials. Domestic terrorism showed an analogous but faster progression among target types, where attacks on private parties overtook those against officials by 1981 with an ever-growing gap thereafter (Gaibulloev, Sandler & Santifort, 2012: 141). Another disturbing trend is that terrorists directed more of their attacks against people than against property after the early 1990s (Brandt & Sandler, 2010, 2012). This is probably due to two drivers: (i) people are a softer target, and (ii) religious fundamentalist terrorists seek greater numbers of casualties. These and other trends in terrorist attacks can inform policymakers on how to allocate their limited counterterrorism resources.

Economic consequences of terrorism

The total direct and indirect costs of the four hijackings on 11 September 2001 have been estimated to be $80 to

Figure 3.

Cumulative number of transnational terrorist incidents by target type, 1968–2010

$90 billion when the gross domestic product (GDP) of the United States in 2001 was about $10 trillion (Kunreuther, Michel-Kerjan & Porter, 2003). The direct costs of terrorism involve the immediate losses associated with a terrorist attack and encompass the value of damaged structures, lives lost, injuries sustained, lost wages, destroyed goods, cleanup, and reduced commerce. Indirect or secondary costs of terrorism include attack-related subsequent expenses such as higher insurance premiums, enhanced security costs, counterterrorism expenses, and lost future commerce. Increased compensation paid to those at high-risk locations may also constitute an indirect cost of terrorism. The distinction between direct and indirect costs of terrorism is not easy to draw. Unlike 9/11, most terrorist attacks have modest economic costs (Enders & Sandler, 2012). Even large-scale attacks have small losses as a percentage of GDP. For example, the costs of the Madrid commuter train bombings on 11 March 2004 were estimated to be $335.1 million or about 0.03% of Spain’s GDP at the time (Busea et al., 2007).

Economic consequences of terrorism can be at the macroeconomic level in terms of reduced GDP or lost GDP growth. Macroeconomic aggregates – consumption, investment, and government expenditures – may also be affected by terrorist attacks. A country that experiences a significant terrorist campaign may lose investment owing to a higher perceived risk on the part of the investors (Blomberg, Hess & Orphanides, 2004). Moreover, such investors may be expected to transfer their funds to other countries with similar rates of returns but lower risks of terrorism. The need for more government-supported counterterrorism measures may crowd out public and private investments owing to higher taxes. There may also be microeconomic level impacts from terrorism as some sectors or locations are at greater peril from terrorist attacks. In a time-series analysis, Enders, Sandler & Parise (1992) showed that terrorist attacks directed at the Greek tourist industry cost Greece 23.4% of its annual tourism revenue for 1988. Another wave of terrorist attacks in Austria during 1985–87 cost Austria 40.7% of its annual tourist revenues for 1988 (also, see Drakos & Kutan, 2003).

There are some general economic principles that can be drawn from the voluminous literature on the economic consequences of terrorism. First, rich, diversified countries are generally able to withstand their annual terrorist attacks with few macroeconomic impacts. This follows because economic activities move from terrorism-prone sectors (provinces) to safer sectors (provinces) within the country (Sandler & Enders, 2004). This first principle is consistent with the observation that most terrorist incidents cause few casualties (i.e. one death on average) and modest property damage based on ITERATE. Advanced economies can offset the effects of terrorism with well-directed fiscal and monetary policies, as the USA did following 9/11 (Enders & Sandler, 2012: 296–298). Second, small terrorism-plagued economies – e.g. Israel, Colombia, and the Basque Country in Spain – will experience significant losses of GDP amounting to upwards of 10% of GDP during sustained terrorist campaigns (e.g. Abadie & Gardeazabal, 2003). Third, small developing countries are apt to show detrimental economic effects from terrorism (Enders & Sandler, 2012; Gaibulloev & Sandler, 2011). This is likely true of failed states that host terrorist groups, because such groups will scare away foreign direct investment (FDI), which is an important source of savings and, hence, growth. If, however, a developing country has only a modest amount of terrorism, then the country should weather these attacks with little economic stress. Fourth, most cross-sectional and panel studies of regional aggregates found that terrorism caused a small, but significant, negative impact on per capita GDP growth (Enders & Sandler, 2012: 298–304). In Blomberg, Hess & Orphanides (2004), the fall in this growth rate was about 0.05% per year for an average country in their worldwide sample (also see Gaibulloev & Sandler, 2011). Fifth, even though terrorism-prone sectors may sustain substantial economic losses, these losses are small relative to the entire economy. This follows because the tourism or FDI-sector is generally small relative to overall GDP. Sixth, the immediate costs of most terrorist attacks are localized, like crime, which can be ameliorated through the transference of economic activities to safer locations. Seventh, major stock markets have shown little reaction to large-scale terrorist incidents, with the exception of 9/11 when world stock indices fell greatly but recovered lost value in 30–40 days (Chen & Siems, 2004). Typically, terrorist attacks affect the stock value of specific companies for a relatively short period of time. For example, Drakos (2004: Table 3) estimated that airline stocks had recovered their full value in less than a year after 9/11, which had an unprecedented effect on these stocks (also see Chen & Siems, 2004).

Effectiveness of counterterrorism policies

Since the Landes (1978) study, there has been a research interest in measuring the effectiveness of specific counterterrorism policies. Even though technological barriers can inhibit a specific type of terrorist incident, such barriers may have unintended consequences as they cause transference of attacks. Transference is an unintended policy-induced change in terrorist behavior, which includes displacement of the mode or venue of attack owing to target hardening. For instance, metal detectors made skyjackings more costly, while they made other kinds of hostage taking relatively less costly. As a consequence, the number of skyjackings fell as kidnappings became more frequent (Enders & Sandler, 1993). In fact, the 1973 installation of metal detectors was associated with an increase in terrorist incidents with casualties (Enders & Sandler, 2012). To truly investigate the effectiveness of such barriers, a researcher must ascertain the impact of the intervention on substitute and complementary modes of attack. In the case of metal detectors, embassy incursions later went down as barriers were installed to protect the compounds (Enders & Sandler, 2012: 166). Landes (1978) did not allow for these other impacts. To judge the effectiveness of an intervention, one would need to know the costs of the action and its resulting benefit. The latter is difficult to determine because it involves a counterfactual – the number of attacks in the absence of the intervention.

Transference is a key consideration for evaluating most counterterrorism actions. For example, enhancing embassy security induced terrorists to assassinate officials outside the compound (Enders & Sandler, 1993). Actions to secure, say, Americans on US soil through homeland security resulted in more attacks against Americans on foreign soil (Enders & Sandler, 2006). Before and after 9/11, about 40% of transnational terrorist attacks were directed against US interests even though very few of these attacks occurred on US soil after 9/11 (Enders & Sandler, 2012). Apparently, terrorists weighed changing risks, as rational action predicts, when they chose where and how to attack.

Another counterterrorism disappointment is the apparent lack of effectiveness of UN resolutions and conventions. Some researchers saw great promise in these treaties (Wilkinson, 1986). Intervention vector autoregression studies showed that past UN resolutions and conventions had no measurable effect on particular outlawed modes of attack (e.g. bombings, skyjackings, or blowing up planes) (Enders & Sandler, 1993, 2012). That is, the pre-intervention mean of the particular outlawed attack mode did not differ from the post-intervention mean after the treaty was ratified. This is not surprising because these treaties have no enforcement mechanism and only a few noncompliant states can undo the gains from compliant states.

Time-series studies also investigated the counterterrorism effectiveness of retaliatory raids, such as the US bombing raid in 1986 on Libya, following the La Belle discotheque bombing in Berlin. Enders & Sandler (1993) found an intertemporal substitution, where terrorists moved attacks planned for the future to the present to protest the raid. As a consequence, terrorist attacks went up following the raid and declined months later as terrorists had to replenish spent resources. The overall level of terrorist attacks did not change in the near term. Brophy-Baermann & Conybeare (1994) discovered the same pattern for Israeli retaliatory raids on the Palestinian terrorists.

Using Palestinian–Israeli data, Dugan & Chenoweth (2012) provided a comparison between the use of repressive deterrence and conciliatory actions against terrorists. The latter actions reward would-be terrorists for refraining from violence. Dugan & Chenoweth’s (2012) parametric and nonparametric empirical analyses showed that indiscriminate repressive actions resulted in more terrorist attacks, whereas indiscriminate conciliatory actions resulted in fewer terrorist attacks. These results are in agreement with past studies about repression-induced backlash attacks (see e.g. Bloom, 2005; Rosendorff & Sandler, 2004; Siqueira & Sandler, 2007).

Theoretical work indicated that governments have more success addressing domestic than transnational terrorism, because a national government can internalize all externalities for potential domestic targets (Enders & Sandler, 2012; Sandler, 2005). Moreover, other nations cannot free ride on its proactive efforts to eradicate a domestic terrorist group. When the French eradicated Action Directe, no other countries derived benefits because the group did not attack the interests of other countries. For a common transnational terrorist threat, sovereign governments generally overspend on defensive measures to shift attacks abroad, while these governments underspend on proactive (offensive) measures in order to free ride on the efforts of other governments (Arce & Sandler, 2005; Sandler & Siqueira, 2006). Thus, defensive measures result in negative transnational externalities,⁴ while proactive measures lead to positive transnational externalities. This overspending on defense may be offset for countries having significant assets abroad. If US homeland security shifts terrorist attacks to the Middle East or Asia where US citizens are killed, the incentive to shift attacks overseas is curtailed. Instead, the USA may be induced to take actions to eliminate the particular terrorist threat altogether. Proactive responses against a terrorist threat may unleash backlash if terrorist supporters view such actions as excessive, thereby resulting in new recruits. When this occurs, the positive benefits conveyed to other countries from the actions are reduced by the backlash costs, so that underspending is curtailed.

Causes and roots of terrorism

Insofar as terrorism has myriad root causes, it is not surprising that the empirical literature has come to little consensus on the root cause of domestic and transnational terrorism for various sample countries. Economic discrimination, religious persecution, nationalist/separatist motives, religious fundamentalism, political ideologies, and other grounds may erupt in terrorism by marginalized groups (Wilkinson, 1986). After 9/11, there was a popular belief that poverty led to transnational terrorism; however, Krueger & Maleckova (2003) uncovered little relationship between the lack of market opportunities and terrorism (also see Piazza, 2006). Opinion polls of Palestinians in the West Bank and the Gaza Strip showed that support for terrorist attacks did not decrease with income or educational level. For a cross-section sample of countries, Krueger & Maleckova (2003) found that low income was not associated with more terrorism when civil and political liberties were controlled. These authors concluded that it is the lack of these liberties, not poverty, that caused terrorism. In a provocative study, Enders & Hoover (2012) argued for a nonlinear relationship between income and terrorism in which middle income is more conducive to terrorism (also see de la Calle & Sánchez-Cuenca, 2012). In poor countries, the population is more focused on day-to-day survival, while in rich countries there are fewer grievances to fuel terrorism.

Given the myriad contradictory findings on the root causes of terrorism, there was a need for an article that tried to sort out the drivers of these differences. To do so, Gassebner & Luechinger (2011) applied extreme bound analysis to identify robust determinants or causes of terrorism. They used ITERATE, GTD, and RAND event datasets. In addition, these authors performed different panel runs for the victim, venue, and perpetrator countries. Key determinants of the number of terrorist attacks differed by the dataset and country identifier. There is more commonality among robust variables for the victim and venue countries than for the perpetrator countries. For the victim country, GDP per capita, population, wars, and religious and ethnic tensions are positive determinants of terrorist attacks, while economic freedoms and physical integrity rights (i.e. absence of human rights abuse) are negative determinants of these attacks. For the venue country, these authors found that economic freedom, physical integrity rights, and law and order are negative influences on terrorism, while population, military spending, involvement in wars, foreign portfolio investment, and political proximity to the United States (in terms of the UN General Assembly voting) are positive influences on terrorism. For panel runs based on the perpetrators’ country of origin, inhibitors of terrorism include economic freedom, physical integrity rights, and the number of telephone mainlines, whereas promoters of terrorism include civil wars and centrist governments. This study is noteworthy because it showed that the data source and the country viewpoint can influence which determinants of terrorism are significant. This realization is consistent with the bewildering number of articles that claimed to find different root causes for terrorism.

Surely, the causes of domestic and transnational terrorism will differ. The roots of domestic terrorism will naturally be country specific, so that large-N cross-sectional or panel studies will provide an average picture that may be a poor indicator for individual countries. In contrast, such an average view may be more appropriate for the cause of transnational terrorism where terrorist groups may be dominated by common grievances, such as left-wing ideologies during the 1970s and 1980s or religious fundamentalism after the start of the 1990s.

Dilemma of liberal democracies

Wilkinson (1986) raised an important terrorism dilemma that plagues some liberal democracies, whose legitimacy rests on the protection of their citizens’ lives and property. His concern is that terrorists can force these liberal democracies into a double blind. If a liberal democracy responds too passively to the terrorist threat and appears unable to guarantee its citizens’ safety, then the government loses its legitimacy and may be voted out of office. If, instead, the government reacts too harshly to the terrorist threat, then the government’s overreaction may compromise democratic principles, thereby making the government less popular. In the process, the terrorists may gain supporters. This dilemma and its dynamics may induce terrorists to locate attacks in liberal democracies rather than in autocracies. Moreover, these democracies may unwittingly facilitate terrorism through executive constraints on power, freedom of association, freedom of speech, right to privacy, press freedoms, a target-rich environment, and other considerations (Hoffman, 2006; Wilkinson, 1986). Starting with Eubank & Weinberg (1994), empirical researchers have been concerned whether democracies are more likely to host terrorist groups than autocracies. Based on an odds ratio test, Eubank & Weinberg (1994) found that terrorist groups are more likely present in a democracy as compared to an autocracy. These authors also discovered that fledgling and transition democracies are vulnerable to terrorism.

In a follow-up article, Li (2005) applied negative binomial regressions to account for the count nature of the dependent variable – i.e. the number of transnational terrorist events. For a sample of 119 countries during 1975–97, Li (2005) found that democratic participation reduced transnational terrorist attacks (also see Eyerman, 1998), while democratic governments’ executive constraints increased transnational terrorist attacks. Numerous subsequent studies investigated the relationship between democracy and terrorism. Based on RAND data, Savun & Phillips (2009) demonstrated that democratic principles positively affected the amount of transnational terrorism, but did not affect the amount of domestic terrorism. As these authors delved further into the cause of transnational terrorism, they found that democracies that engaged in foreign policy crises with other countries attracted more transnational terrorism. Moreover, alliance ties with the United States and interventions in foreign civil wars also resulted in the country experiencing more transnational terrorist attacks. Thus, democratic principles per se did not generate such attacks.

The relationship between democracy and terrorism remains an active area of research. For example, Chenoweth (2010) argued that intergroup competition is a driver of transnational terrorist incidents in democracies as terrorist groups vie for supporters. Chenoweth’s unit of analysis was a cross-national longitudinal study of 119 countries for 1975–97, based on ITERATE data. She did not investigate whether intergroup competition was also more conducive to domestic terrorism. Chenoweth (2010) also presented evidence that political competition resulted in more terrorist groups in the country. In a recent panel analysis, Young & Dugan (2011) showed that the presence of veto players (i.e. entities that can block policy changes) was positively related to homegrown and all fatal attacks based on GTD event data.

Major new developments in the study of terrorism

I now present a few areas where there is much current and, I anticipate, future interest. GTD data provide a chance to deconstruct terrorism into domestic and transnational incidents in order to test a host of hypotheses. This opportunity was afforded once Enders, Sandler & Gaibulloev (2011) partitioned GTD into its two component types of terrorism. Table I indicates hypothesized and tested differences between domestic and transnational terrorism. For instance, Piazza (2011) found that economic discrimination was a clear determinant of domestic terrorism; however, this discrimination was not

Table I.

Reasons for distinguishing between domestic and transnational terrorism

Causes of each type of terrorism may differ – e.g. economic discrimination for domestic terrorism (Piazza, 2011) and foreign policy decisions for transnational terrorism (Savun & Phillips, 2009).

The two types of terrorism may display different dynamic responses – e.g. domestic terrorism may Granger-cause transnational terrorism, but transnational terrorism may not Granger-cause domestic terrorism (Enders, Sandler & Gaibulloev, 2011).

The effectiveness of counterterrorist policies may differ between the two types of terrorism because sovereign governments may not account for the associated transnational externalities of their countermeasures against transnational terrorism. For domestic terrorism, there are no transnational externalities. (Not tested.)

Counterterrorist costs differ between the two kinds of terrorism – e.g. the costs to fight transnational terrorism are high as a country projects its power beyond its borders to a terrorist group’s safe haven. (Not tested.)

Transnational terrorism may require international cooperation, which is generally not needed for domestic terrorism. (Not tested.)

The macroeconomic consequences may differ – i.e. transnational terrorism may affect growth and other macroeconomic aggregates in a different manner than domestic terrorism (Gaibulloev & Sandler, 2008, 2011).

Terrorism’s impact on various economic sectors may differ – e.g. transnational terrorism affects tourism, FDI, and trade (Enders & Sandler, 2012), while domestic terrorism impacts provinces within the country (Abadie & Gardeazabal, 2003).

Domestic terrorist attacks are likely tied to internal wars, while transnational terrorist attacks are likely tied to interstate wars. (Not tested.)

The survival of terrorist groups may hinge on their mix of transnational and domestic terrorist attacks owing to the enhanced risks posed by transnational terrorist attacks (Blomberg, Gaibulloev & Sandler, 2011).

Homegrown terrorism is more linked than transnational terrorism to the presence of veto players (Young & Dugan, 2011).

Remittances are more supportive of domestic than of transnational terrorism (Mascarenhas & Sandler, 2013).

Articles testing the hypotheses are indicated.

an unequivocal determinant of transnational terrorism. This follows because attacks against foreign venues or assets may be motivated by foreign policy decisions and other perceived grievances. Using country-year count data on terrorist attacks from RAND, Savun & Phillips (2009) found that foreign policy was a significant determinant of transnational, but not of domestic, terrorism in democracies. Additional hypotheses, along with the articles supporting them, are displayed in Table I. As indicated, some suggested hypotheses are not yet tested. Many other hypotheses, distinguishing domestic and transnational terrorism, are tested, but space does not permit them to be presented.

Following 9/11, there has been a large interest in applying game theory to the study of terrorism. Game theory – or the study of strategic rational choice – is an appropriate tool to capture the strategic interactions among various agent pairings – e.g. two governments targeted by the same terrorist group; a government and its adversarial terrorist group; rival terrorist groups; the government, the terrorists, and the media; or the government and factions within a terrorist group. Table II presents a few novel findings from this literature. I favor findings that are either provocative or counter-intuitive. Most game-theoretic treatments involved counterterrorism applications, as implied by Table II. To date, there are not many tests of these propositions, primarily because there are limited data on counterterrorism actions of governments – e.g. we do not have data on how governments allocate defensive resources. Thus, empirical support in articles relies on a few specific examples. There are a few exceptions, where the propositions are empirically tested. For example, Benmelech, Berrebi & Klor (2012) presented empirical support for Bueno de Mesquita’s (2005) quality of terrorism hypothesis. Based on Jones & Libicki (2008) data on terrorist groups, Bapat (2011) provided cross-sectional support for his moral hazard argument about US military aid to host countries.

The game-theoretic analysis of terrorism provides a number of insights. First, strategic behavior may induce allies – e.g. commonly targeted governments – to work at cross-purposes (Bueno de Mesquita, 2007; Sandler & Siqueira, 2006). Second, adversaries may also work at cross-purposes as each tries to match the response of their counterpart without gaining any ground (Arce & Sandler, 2005). Third, information plays an important role since adversaries are often ill-informed about their counterparts’ resources or preferences (Kydd & Walter, 2002). This lack of information makes signaling and screening strategies important (Lapan & Sandler, 1993).

Table II.

Some novel game-theoretic results

Increased information by the governments on the targeting preferences of the common terrorist threat will worsen defensive allocations as governments are more effective in transferring attacks abroad (Sandler & Lapan, 1988).

Whether terrorists are able to sabotage peace agreements depends on the level of trust between the government and the moderate terrorists. The ability to sabotage also hinges on the strength of these moderates and the orientation of the government (Kydd & Walter, 2002).

Terrorist organizations screen volunteers for quality, so that they attract the most capable operatives. Reduced economic opportunities allow these groups to obtain better recruits (Bueno de Mesquita, 2005).

Among two targeted countries, leadership will exacerbate the undersupply of proactive measures, but it will ameliorate the oversupply of defensive actions. The net effect of leadership on the efficient supply of counterterrorism depends on these opposing influences (Sandler & Siqueira, 2006).

Within a country, a centralized allocation is more efficient than a noncentralized allocation of defensive measures (Bier, Oliveros & Samuelson, 2007).

Generally, defensive allocation decisions should be made public, so that terrorists are guided to low-valued targets (Bier, Oliveros & Samuleson, 2007).

Governments overspend on tactic-specific observable countermeasures, while they underspend on unobservable covert operations (Bueno de Mesquita, 2007).

The attacker’s and defender’s efforts may be strategic complements or strategic substitutes. For complements, the attacker’s efforts rise with increases in the defender’s efforts, so that the defender may not improve its security despite its extra efforts (Zhuang & Bier, 2007).

When a defender must protect N locations from terrorist attacks, the solution algorithm involves allocating resources among sites to minimize the expected gain of the terrorists even though the game is not zero sum (Powell, 2007).

The choice between proactive and defensive countermeasures by two targeted countries hinges on three considerations: (i) the relative cost comparisons of the policies; (ii) the foreign interests of the countries; and (iii) the targeting risks of the countries. A high-cost defender is more motivated to provide a proactive response, because it suffers greater costs (Bandyopadhyay & Sandler, 2011).

US military aid to countries hosting a terrorist group creates a moral hazard problem for the recipient country that lengthens the resident group’s lifetime (Bapat, 2011).

Fourth, by taking a long-term viewpoint, terrorist groups are more likely than targeted governments to work in conjunction, thereby resulting in terrorist networks (Sandler, 2005). In contrast, democratic governments cannot see past the next election, so that agreements with other governments are not so compelling because the next elected government may not abide by past agreements (Sandler, 2005). Fifth, as more agents have been added to a game model, theorists have resorted to multistage games, where each stage typically contains two active agents.

In addition, game theory articles investigated bargaining between hostage-taking terrorists and governments (see Sandler, Tschirhart & Cauley, 1983). For example, Lapan & Sandler (1988) studied the practicality of a policy that commits governments never to concede to hostage takers’ demands. The conventional support of the no-concession policy hinges on the notion that if terrorists know beforehand that they have nothing to gain, then they will never abduct hostages. The empirical record shows that governments do offer concessions in violation of their pledge – e.g. the Irangate scandal involving the Reagan administration. Lapan & Sandler (1988) showed that these failures to abide by a stated no-concession policy hinge on a time-inconsistency concern, when the perceived short-run gain from concessions appears to be greater than the long-run consequences owing to the value of a given hostage – e.g. CIA agent William Buckley kidnapped on 16 March 1984 in Beirut, who precipitated the Irangate scandal. Many game-theoretic bargaining propositions have been tested by Atkinson, Sandler & Tschirhart (1987) and Gaibulloev & Sandler (2009). The latter article showed that governments have abided by their no-concession pledge more often in recent years.

Many of the original empirical studies of terrorism involved time-series analysis (see e.g. Enders, 2007). As count models, such as Poisson and zero-inflated negative binomial, were developed, these models replaced those based on a normal distribution. After 9/11, there are many more cross-sectional and panel studies on such topics as suicide terrorism (Benmelech, Berrebi & Klor, 2012), the causes of terrorism (Piazza, 2011), effectiveness of counterterrorism (Dugan & Chenoweth, 2012), and economic consequences of terrorism (Gaibulloev & Sandler, 2008, 2011).

Another recent research development concerned the study of networked terrorists. In the late 1960s and the 1970s, terrorist networks cooperated on many levels, including training, intelligence, safe haven, financial support, logistical help, weapon acquisition, and the exchange of personnel (Hoffman, 2006). The European left-wing terrorists maintained ties with the Palestinian groups (Alexander & Pluchinsky, 1992). In recent times, the Al-Qaeda network operates in upward of 60 countries and stages its attacks worldwide. These terrorist networks allow tied groups to increase their effectiveness in utilizing limited resources compared to their powerful government adversaries.

One approach to the study of terrorist networks is to design counterterrorist policies to do the most damage to compromise linkages between the terrorists (Farley, 2003). This approach relies on graph theory and assumes that policy designers have knowledge about the network’s configuration. Moreover, it assumes a static configuration, where the terrorists do not alter links in anticipation of policymakers’ goal to disrupt the network. The focus on networks teaches us that terrorists must trade off connectivity against vulnerability (Enders & Jindapon, 2010). In an all-channel network, every terrorist is linked to every other terrorist, which facilitates complex operations as information can be exchanged readily. In such a network, a single compromised terrorist can provide devastating intelligence on the rest of the network. To avoid this vulnerability, terrorist networks rely on rather small chains of terrorists who are loosely linked to the rest of the network (Enders & Sandler, 2012: 249–256). The resulting downside is that planned attacks must be rather simple.

Networks are also relevant in a game-theoretic setting to investigate how a global terrorist organization (GTO) delegates local terrorist groups in the field (Siqueira & Sandler, 2010). This delegation decision is based on the GTO’s perceived orientation of the targeted government and the fervor of the local supporters for the terrorist group’s goals. If, for example, the targeted government is weak with no stomach to fight the terrorists (i.e. it does not increase counterterrorism as attacks increase), and if the local supporters appear steadfast, then the GTO would delegate to a local group, whose views are more extreme than its own. If the targeted government and/or local supporters later changed their orientation, then the GTO may come to regret its delegation decision with little chance to recall the deployed group. This was the case with Al-Qaeda’s deployment of al-Zarqawi to Iraq after he bombed three Jordanian hotels with high Muslim death tolls.

Because many terrorist groups reside in developing countries with limited means to eradicate them, the use of foreign aid to calm grievances and to bolster proactive counterterrorism efforts against a resident group is now an important research question (e.g. Azam & Thelan, 2010). Relevant issues include the interplay between a donor country, which is focused on defensive measures at home against an attack launched from abroad, and the recipient country, which is focused on proactive measures to eliminate the resident terrorist group (Bandyopadhyay, Sandler & Younas, 2011). This interplay is strategically tied to the donor’s aid package, which is used not only to bolster the recipient country’s well-being, but also to finance its proactive efforts. Other germane considerations involve the form of the foreign assistance and its influence on the regime stability of the aid recipient. Too much aid tied to counterterrorism may lose the recipient regime its popular support (Bandyopadhyay, Sandler & Younas, 2011). In a recent empirical study, Young & Findley (2011) showed that foreign aid reduced transnational terrorism when targeted toward education, health, civil society, and conflict prevention. Based on panel analyses of 22 donor countries, Dreher & Fuchs (2011) showed that foreign aid increased after 9/11; however, this aid did not necessarily flow to the countries with the most terrorist attacks. Surprisingly, these authors did not find that aid flowed to the perpetrators’ host countries. Many other aid-related issues for the study of terrorism remain.

Directions for future research and concluding remarks

What is the true impact of terrorism on growth? For regional aggregates, the literature’s consensus is that terrorism has a small, but significant, negative impact on per capita GDP growth (see e.g. Blomberg, Hess & Orphanides, 2004). There are, however, grounds for believing that terrorism typically has no effect on economic growth – e.g. most countries experience few attacks with little loss of life or property; much of terrorism’s impact is localized; and transference of economic activities can limit economic consequences. To date, the studies of this growth impact ignores biases – e.g. Nickell bias stemming from lagged per capita income on the right-hand side of the growth equation. Another crucial bias comes from cross-country dependence which causes an endogeneity bias. When accounting for these biases, Gaibulloev, Sandler & Sul (2013) found no influence of terrorism on economic growth for a host of regional samples. More work is needed because losses in economic growth are used to gauge offsetting fiscal and monetary policies.

There is also a need for more spatial analysis of terrorist incidents and civil wars. Recently, geo-referenced GTD data were used by Findley & Young (2012) to investigate the spatial and temporal relationship between terrorist incidents and civil wars. Spatial analysis can indicate the spread of terrorist campaigns across countries and regions. The analysis can also identify where vulnerabilities are the greatest within countries, and how terrorist attacks respond spatially to counterterrorism measures.

There are additional issues concerning counterterrorism spending that require investigation. How much is enough in terms of defensive spending? That is, have past counterterrorism measures been effective and worth the cost? Since its inception, the budget of US DHS has risen by 9% annually (Enders & Sandler, 2012: 328). To address such questions requires an unobservable counterfactual, in which benefits in terms of inhibited attacks are ascertained. The cost side of counterterrorism is readily observable. There is only a single study that attempts to compute a counterfactual for post-9/11 counterterrorism spending (Sandler, Arce & Enders, 2011). Though this study is imperfect, these authors found only about a ten cents return on a dollar from post-9/11 defensive spending.

Datasets, such as ITERATE, GTD, and RAND, attribute attacks to terrorist groups for the last 40 years. Even though less than half of the recorded attacks are tied to specific groups, there are now sufficient data to investigate the determinants of terrorist group survival. The pioneering article using survival (hazard) analysis is by Blomberg, Engel & Sawyer (2010), with its focus on ‘one hit wonders’ or groups that struck just once. Their study investigated how economic, political, and geographical considerations in the venue country influenced terrorist group longevity. A follow-up study by Blomberg, Gaibulloev & Sandler (2011) included terrorist groups’ ideologies and tactics (e.g. attack diversity and share of transnational attacks). This second study relied on economic, political, and geographical variables but in the terrorist group’s base country. Future analyses need to go a step further to condition the groups’ failure to a cause – i.e. loss of a leader, dissent within the group, accomplishment of its political goal, or defeat by the authorities. Data in a study by Jones & Libicki (2008) can be used to explain specific causes of group failure (Carter, 2012).

Suicide terrorism requires further analysis. Benmelech, Berrebi & Klor (2012) established that poor economic conditions may result in terrorist groups recruiting more capable and better educated suicide bombers to attack Israeli targets. As a result, the ‘quality’ of attacks increased in terms of carnage because more qualified suicide bombers can wreak more casualties. This important work should be extended to a global sample.

Yet another issue for future research is the study of the factors that promote international cooperation in the fight against international terrorism. For example, INTERPOL can coordinate arrest efforts among countries. The aftermath of 9/11 showed that, under dire threat, nations can cooperate against a common terrorist enemy. For counterterrorism, the role of partial cooperation needs to be investigated.

A final direction for future research involves the development of efforts to counter the growing threat of cyberterrorism in an increasingly electronic-dependent world.

This article demonstrates that the analytical study of terrorism is a very active research area that has much policy relevance. Event datasets facilitated empirical tests of theoretical propositions, but underscored a real deficiency – i.e. the need for data on governments’ counterterrorism actions. There are a few country-specific counterterrorism datasets (see e.g. Dugan & Chenoweth, 2012), but no global counterterrorism dataset. This article also highlights select areas of major new developments, as well as future directions for research.

Footnotes

Replication data

The raw data for the figures along with the figures are in an Excel file that can be found at

Acknowledgements

The author has profited from helpful comments provided by four anonymous reviewers and the two editors of the special issue.

References

Abadie

Alberto

Gardeazabal

Javier

(2003) The economic costs of conflict: A case study of the Basque Country. American Economic Review 93(1): 113–132.

Abrahms

Max

(2008) What terrorists really want: Terrorist motives and counterterrorism strategy. International Security 32(4): 78–105.

Alexander

Yonah

Pluchinsky

Dennis

(1992) Europe’s Red Terrorists: The Fighting Communist Organizations. London: Frank Cass.

Arce

Daniel G

Sandler

Todd

(2005) Counterterrorism: A game-theoretical analysis. Journal of Conflict Resolution 49(2): 183–200.

Atkinson

Scott E

Sandler

Todd

Tschirhart

John

(1987) Terrorism in a bargaining framework. Journal of Law Economics 30(1): 1–21.

Azam

Jean-Paul

Thelen

Véronique

(2010) Foreign aid versus military intervention in the war on terror. Journal of Conflict Resolution 54(2): 237–261.

Bandyopadhyay

Subhayu

Sandler

Todd

(2011) The interplay between preemptive and defensive counterterrorism measures: A two-stage game. Economica 78(3): 546–564.

Bandyopadhyay

Subhayu

Sandler

Todd

Younas

Javed

(2011) Foreign aid as counterterrorism policy. Oxford Economic Papers 63(3): 423–447.

Bapat

Navin A

(2011) Transnational terrorism, US military aid, and the incentive to misrepresent. Journal of Peace Research 48(3): 303–318.

10.

Benmelech

Efraim

Berrebi

Claude

Klor

Esteban F

(2012) Economic conditions and the quality of suicide terrorism. Journal of Politics 74(1): 113–128.

11.

Bier

Vicki

Oliveros

Santiago

Samuelson

Larry

(2007) Choosing what to protect: Strategic defensive allocation against an unknown attacker. Journal of Public Economic Theory 9(4): 563–587.

12.

Blomberg

S Brock

Engel

Rozlyn C

Sawyer

Reid

(2010) On the duration and sustainability of transnational terrorist organizations. Journal of Conflict Resolution 54(2): 303–330.

13.

Blomberg

S Brock

Gaibulloev

Khusrav

Sandler

Todd

(2011) Terrorist group survival: Ideology, tactics, and base of operations. Public Choice 149 (3–4): 441–463.

14.

Blomberg

S Brock

Hess

Gregory D

Orphanides

Athanasios

(2004) The macroeconomic consequences of terrorism. Journal of Monetary Economics 51(5): 1007–1032.

15.

Bloom

Mia

(2005) Dying to Kill: The Allure of Suicide Terror. New York: Columbia.

16.

Brandt

Patrick T

Sandler

Todd

(2010) What do transnational terrorists target? Has it changed? Are we safer? Journal of Conflict Resolution 54(2): 214–236.

17.

Brandt

Patrick T

Sandler

Todd

(2012) A Bayesian Poisson vector autoregression model. Political Analysis 20(3): 292–315.

18.

Brophy-Baermann

Bryan

Conybeare

John AC

(1994) Retaliating against terrorism: Rational expectations and the optimality of rules versus discretion. American Journal of Political Science 38(1): 196–210.

19.

Bueno de Mesquita

Ethan

(2005) The quality of terror. American Journal of Political Science 49(3): 515–530.

20.

Bueno de Mesquita

Ethan

(2007) Politics and the suboptimal provision of counterterror. International Organization 61(1): 9–36.

21.

Busea

Mikel

Valiño

Aurelia

Heijs

Joost

Baumort

Thomas

Gomez

Javier Gonzalez

(2007) The economic cost of March 11: Measuring the direct economic cost of the terrorist attack on March 11, 2004 in Madrid. Terrorism and Political Violence 19(4): 489–509.

22.

Carter

David

(2012) A blessing or curse? State support for terrorist groups. International Organization 66(1): 129–151.

23.

Chen

Andrew H

Siems

Thomas F

(2004) The effects of terrorism on global capital markets. European Journal of Political Economy 20(2): 249–266.

24.

Chenoweth

Erica

(2010) Democratic competition and terrorist activity. Journal of Politics 77(1): 16–30.

25.

Crenshaw

Martha

(1981) The causes of terrorism. Comparative Politics 13(4): 379–399.

26.

Luis

De la Calle

Sánchez-Cuenca

Ignacio

(2012) Rebels without a territory: An analysis of nonterritorial conflicts in the world, 1970–1997. Journal of Conflict Resolution 56(4): 580–603.

27.

Drakos

Konstantinos

(2004) Terrorism-induced structural shifts in financial risk: Airline stocks in the aftermath of the September 11th terror attacks. European Journal of Political Economy 20(2): 435–466.

28.

Drakos

Konstantinos

Kutan

Ali M

(2003) Regional effects of terrorism on tourism in three Mediterranean countries. Journal of Conflict Resolution 47(5): 621–641.

29.

Dreher

Axel

Fuchs

Andreas

(2011) Does terror increase aid? Public Choice 149 (3–4): 337–363.

30.

Dugan

Laura

Chenoweth

Erica

(2012) Moving beyond deterrence: The effectiveness of raising the expected utility of abstaining from terrorism in Israel. American Sociological Review 77(4): 597–624.

31.

Enders

Walter

(2007) Terrorism: An empirical analysis. In: Sandler

Todd

Hartley

Keith

(eds) Handbook of Defense Economics, Vol. 2. Amsterdam: North-Holland, 815–866.

32.

Enders

Walter

Hoover

Gary A

(2012) The nonlinear relationship between terrorism and poverty. American Economic Review 102(3): 267–272.

33.

Enders

Walter

Jindapon

Paan

(2010) Network externalities and the structure of terror networks. Journal of Conflict Resolution 54(2): 262–280.

34.

Enders

Walter

Sandler

Todd

(1993) The effectiveness of anti-terrorism policies: A vector-autoregression-intervention analysis. American Political Science Review 87(4): 829–844.

35.

Enders

Walter

Sandler

Todd

(2000) Is transnational terrorism becoming more threatening? A time-series investigation. Journal of Conflict Resolution 44(3): 307–332.

36.

Enders

Walter

Sandler

Todd

(2006) Distribution of transnational terrorism among countries by income classes and geography after 9/11. International Studies Quarterly 50(2): 367–393.

37.

Enders

Walter

Sandler

Todd

(2012) The Political Economy of Terrorism. Cambridge: Cambridge University Press.

38.

Enders

Walter

Sandler

Todd

Gaibulloev

Khusrav

(2011) Domestic versus transnational terrorism: Data, decomposition, and dynamics. Journal of Peace Research 48(3): 319–337.

39.

Enders

Walter

Sandler

Todd

Parise

Gerald F

(1992) An econometric analysis of the impact of terrorism on tourism. Kyklos 45(4): 531–554.

40.

Eubank

William L

Weinberg

Leonard B

(1994) Does democracy encourage terrorism? Terrorism and Political Violence 6(4): 417–435.

41.

Eyerman

Joe

(1998) Terrorism and democratic states: Soft targets or accessible systems? International Interactions 24(2): 151–170.

42.

Farley

Jonathan D

(2003) Breaking Al Qaida: A mathematical analysis of counterterrorism operations. Studies in Conflict & Terrorism 26(6): 399–411.

43.

Findley

Michael E

Young

Joseph K

(2012) Terrorism and civil war: A spatial and temporal approach to a conceptual problem. Perspectives on Politics 10(2): 285–305.

44.

Gaibulloev

Khusrav

Sandler

Todd

(2008) Growth consequences of terrorism in Western Europe. Kyklos 61(3): 411–424.

45.

Gaibulloev

Khusrav

Sandler

Todd

(2009) Hostage taking: Determinants of terrorist logistical and negotiation success. Journal of Peace Research 46(6): 739–756.

46.

Gaibulloev

Khusrav

Sandler

Todd

(2011) The adverse effect of transnational and domestic terrorism on growth in Africa. Journal of Peace Research 48(3): 355–371.

47.

Gaibulloev

Khusrav

Sandler

Todd

Santifort

Charlinda

(2012) Assessing the evolving threat of terrorism. Global Policy 3(2): 135–144.

48.

Gaibulloev

Khusrav

Sandler

Todd

Sul

Donggyu

(2013) Reevaluating terrorism and growth: Dynamic panel analysis and cross-sectional dependence. Unpublished paper. School of Economic, Political & Policy Sciences, University of Texas at Dallas.

49.

Gassebner

Martin

Luechinger

Simon

(2011) Lock, stock, and barrel: A comprehensive assessment of the determinants of terror. Public Choice 149 (3–4): 235–261.

50.

Gould

Eric D

Klor

Esteban F

(2010) Does terrorism work? Quarterly Journal of Economics 125(4): 1459–1510.

51.

Hoffman

Bruce

(2006) Inside Terrorism , revised edn. New York: Columbia University Press.

52.

Jones

Seth G

Libicki

Martin C

(2008) How Terrorist Groups End: Lessons for Countering al Qa’ida. Santa Monica, CA: RAND.

53.

Krueger

Alan

Maleckova

Jitka

(2003) Education, poverty, and terrorism: Is there a causal connection? Journal of Economic Perspectives 17(4): 119–144.

54.

Kunreuther

Howard

Michel-Kerjan

Erwann

Porter

Beverly

(2003) Assessing managing and financing extreme events: Dealing with terrorism. Working Paper 10179. Cambridge, MA: National Bureau of Economic Research.

55.

Kydd

Andrew

Walter

Barbara F

(2002) Sabotaging the peace: The politics of extremist violence. International Organization 56(2): 263–296.

56.

LaFree

Gary

Dugan

Laura

(2007) Introducing the Global Terrorism Database. Terrorism and Political Violence 19(2): 181–204.

57.

Landes

William M

(1978) An economic study of US aircraft hijackings, 1961–1976. Journal of Law and Economics 21(1): 1–31.

58.

Lapan

Harvey E

Sandler

Todd

(1988) To bargain or not to bargain? That is the question. American Economic Review 78(2): 16–20.

59.

Lapan

Harvey E

Sandler

Todd

(1993) Terrorism and signalling. European Journal of Political Economy 9(3): 383–397.

60.

Quan

(2005) Does democracy promote transnational terrorist incidents? Journal of Conflict Resolution 49(2): 278–297.

61.

Mascarenhas

Raechelle

Sandler

Todd

(2013) Remittances and terrorism: A global analysis. Unpublished paper. School of Economic, Political & Policy Sciences, University of Texas at Dallas.

62.

Mickolus

Edward F

Sandler

Todd

Murdock

Jean M

Flemming

Peter

(2012) International Terrorism: Attributes of Terrorism Events (ITERATE), 1968–2011. Dunn Loring, VA: Vinyard Software.

63.

Pape

Robert A

(2003) The strategic logic of suicide terrorism. American Political Science Review 97(3): 343–361.

64.

Piazza

James A

(2006) Rooted in poverty? Terrorism, poor economic development, and social cleavages. Terrorism and Political Violence 18(1): 159–177.

65.

Piazza

James A

(2011) Poverty, minority economic discrimination and domestic terrorism. Journal of Peace Research 48(3): 339–353.

66.

Powell

Robert

(2007) Defending against terrorist attacks with limited resources. American Political Science Review 101(3): 527–541.

67.

RAND (2012) RAND Database of Worldwide Terrorism Incidents (http://www.rand.org/nsrd/projects/terrorism-incidents.html).

68.

Rapoport

David C

(2004) Modern terror: The four waves. In: Cronin

Audrey K

Ludes

James M

(eds) Attacking Terrorism: Elements of a Grand Strategy. Washington, DC: Georgetown University Press, 46–73.

69.

Rosendorff

B Peter

Sandler

Todd

(2004) Too much of a good thing? The proactive response dilemma. Journal of Conflict Resolution 48(5): 657–671.

70.

Sandler

Todd

(2005) Collective versus unilateral responses to terrorism. Public Choice 124 (1–2): 75–93.

71.

Sandler

Todd

Enders

Walter

(2004) An economic perspective on transnational terrorism. European Journal of Political Economy 20(2): 301–316.

72.

Sandler

Todd

Lapan

Harvey E

(1988) The calculus of dissent: An analysis of terrorists’ choice of targets. Synthese 76(2): 245–261.

73.

Sandler

Todd

Siqueira

Kevin

(2006) Global terrorism: Deterrence versus preemption. Canadian Journal of Economics 50(4): 1370–1387.

74.

Sandler

Todd

Arce

Daniel G

Enders

Walter

(2009) Transnational terrorism. In: Lomberg

Bjørn

(ed.) Global Crises, Global Solutions, 2nd edn. Cambridge: Cambridge University Press, 516–562.

75.

Sandler

Todd

Tschirhart

John

Cauley

Jon

(1983) A theoretical analysis of transnational terrorism. American Political Science Review 77(1): 36–54.

76.

Savun

Burcu

Phillips

Brian J

(2009) Democracy, foreign policy, and terrorism. Journal of Conflict Resolution 53(6): 878–904.

77.

Siqueira

Kevin

Sandler

Todd

(2007) Terrorist backlash, terrorism mitigation, and policy delegation. Journal of Public Economics 91(9): 1800–1815.

78.

Siqueira

Kevin

Sandler

Todd

(2010) Terrorist networks, support, and delegation. Public Choice 142 (1–2): 237–253.

79.

Wilkinson

Paul

(1986) Terrorism and the Liberal State , revised edn. London: Frank Cass.

80.

Young

Joseph K

Dugan

Laura

(2011) Veto players and terror. Journal of Peace Research 48(1): 19–33.

81.

Young

Joseph K

Findley

Michael G

(2011) Can peace be purchased? A sectoral-level analysis of aid’s influence on transnational terrorism. Public Choice 149 (3–4): 365–381.

82.

Zhuang

Jun

Bier

Vicki M

(2007) Balancing terrorism and natural disasters: Defensive strategy with endogenous attacker effort. Operations Research 55(5): 976–991.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB