CREATE Research Archive Published Articles & Papers 2011 An Evaluation of INTERPOL's Cooperative-Based Counterterrorism Linkages Todd Sandler University of Texas at Dallas, [email protected]Daniel G. Arce University of Texas at Dallas, [email protected]Walter Enders University of Alabama - Tuscaloosa, [email protected]Follow this and additional works at: hp://research.create.usc.edu/published_papers is Article is brought to you for free and open access by CREATE Research Archive. It has been accepted for inclusion in Published Articles & Papers by an authorized administrator of CREATE Research Archive. For more information, please contact [email protected]. Recommended Citation Sandler, Todd; Arce, Daniel G.; and Enders, Walter, "An Evaluation of INTERPOL's Cooperative-Based Counterterrorism Linkages" (2011). Published Articles & Papers. Paper 148. hp://research.create.usc.edu/published_papers/148
53
Embed
An Evaluation of INTERPOL's Cooperative-Based ...create.usc.edu/.../anevaluationofinterpolscooperative-basedcounterterrorismli_0.pdfAn Evaluation of INTERPOL’s Cooperative-based
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
CREATE Research Archive
Published Articles & Papers
2011
An Evaluation of INTERPOL's Cooperative-BasedCounterterrorism LinkagesTodd SandlerUniversity of Texas at Dallas, [email protected]
Follow this and additional works at: http://research.create.usc.edu/published_papers
This Article is brought to you for free and open access by CREATE Research Archive. It has been accepted for inclusion in Published Articles & Papersby an authorized administrator of CREATE Research Archive. For more information, please contact [email protected].
Recommended CitationSandler, Todd; Arce, Daniel G.; and Enders, Walter, "An Evaluation of INTERPOL's Cooperative-Based Counterterrorism Linkages"(2011). Published Articles & Papers. Paper 148.http://research.create.usc.edu/published_papers/148
Department of Economics and Finance University of Alabama
College of Business Administration Tuscaloosa, AL 35487 USA
[email protected] Center for Global Collective Action Working Paper No. 10-001
ABSTRACT
This paper evaluates the payback from INTERPOL’s efforts to coordinate proactive
counterterrorism measures by its member countries to arrest terrorists and weaken their
ability to conduct operations. We use INTERPOL arrest data and utilization of
INTERPOL resources by member countries to compute counterfactual benefit measures,
which, when matched with costs, yield benefit-cost ratios. These ratios average around
200 over 12 alternative counterfactual scenarios, so that each dollar of INTERPOL
counterterrorism spending returns around $200. The paper also puts forward a
perspective on benefits derived from INTERPOL’s Stolen and Lost Travel Document
database. INTERPOL provides an inexpensive proactive measure against transnational
terrorism that, unlike military operations, does not result in backlash attacks.
Keywords: INTERPOL; counterterrorism; international cooperation; benefit-cost
analysis; game theory
Forthcoming in Journal of Law and Economics, Vol. 54, February 2011
*Sandler’s research was funded, in part, by the US Department of Homeland Security (DHS) through the Center for Risk and Economic Analysis of Terrorism Events (CREATE) at the University of Southern California, grant number 2007-ST-061-000001. However, any opinions, findings, and conclusions or recommendations are solely those of the authors and do not necessarily reflect the view of the DHS or CREATE. The authors gratefully acknowledge helpful comments from an anonymous referee.
An Evaluation of INTERPOL’s Cooperative-based Counterterrorism Linkages
1. Introduction
Formed in 1923, the International Criminal Police Organization (henceforth, INTERPOL) fosters
transnational police cooperation in the organization’s mission to curb a host of international
criminal activities. Currently, INTERPOL focuses on six crime areas: (i) corruption, (ii)
fugitives, (iii) drugs and organized crime, (iv) public safety and terrorism, (v) trafficking in
human beings, and (vi) financial and high-technology crime (INTERPOL 2008a). To address
these areas, INTERPOL uses its resources and linkages to assist and coordinate the crime
fighting efforts of its member countries. As an independent international organization,
INTERPOL is primarily supported by its 188 member nations through assessed dues based on
ability to pay or income considerations. Well over 80 percent of INTERPOL’s 2008 operating
expense of approximately 54.6 million euros comes from assessed dues, with the remainder
coming from voluntary contributions.1
INTERPOL has assumed a larger role in curbing transnational terrorist attacks that,
through their victims, perpetrators, or implications, impact two or more countries. Terrorist
incidents that begin in one country and end in another (e.g., many skyjackings) are transnational,
as are assassinations or kidnappings of foreigners by a domestic terrorist group (Enders and
Sandler 2006a, pp. 6–7). If, moreover, the perpetrators cross an international border or receive
training and support from abroad, then the terrorist incident is transnational. To direct its
counterterrorism activities, INTERPOL created a Public Safety and Terrorism (PST) sub-
directorate in October 2001 following the September 11, 2001 hijackings (henceforth, 9/11) in
the United States. In so doing, INTERPOL focused more of its crime-fighting resources on
counterterrorism. In particular, INTERPOL assumed a proactive role in fighting transnational
terrorism by bolstering international cooperative linkages among law enforcement agencies,
2
INTERPOL, and its member countries’ National Central Bureaus (NCBs). Some of these
linkages are dependent on technology. The resulting proactive measures are aimed at capturing
terrorists, thwarting planned incidents, and reducing terrorist resources (INTERPOL 2008b).
NCBs consist of skilled law enforcement personnel from the countries concerned and are
controlled by the law enforcement structure of their country; they interface with INTERPOL at
its General Secretariat in Lyon, France and at its six regional bureaus (INTERPOL 2008a).
The primary purpose of this paper is to evaluate the payback from INTERPOL’s efforts
to coordinate proactive counterterrorism measures by its member countries. Such proactive
measures yield positive externalities as the terrorist threat to all at-risk nations is curtailed. The
presence of these positive side benefits means that such proactive measures are likely to be
undersupplied when provided independently by nations (Arce and Sandler 2005; Enders and
Sandler 1995; Sandler and Siqueira 2006). INTERPOL’s actions to link countries after 9/11 in
counterterrorism measures provide potentially large network externalities. Our effort to evaluate
INTERPOL’s antiterrorism contribution requires that its costs be apportioned to the
organization’s counterterrorism activities, since INTERPOL does much more than fight
terrorism. Our study also necessitates a difficult counterfactual calculation as to the number of
transnational terrorist events in the absence of INTERPOL’s actions. We base our
counterfactuals on the utilization of INTERPOL tools and resources – particularly, its facilitated
arrests of terrorist suspects and “hits” (i.e., positive matches) on its Stolen and Lost Travel
Document (SLTD) database. Thus, we link INTERPOL’s resources – its secure global police
communication link, its terrorism-related databases, its police support services, and its training
workshops – to the arrests of suspected terrorists that stemmed from INTERPOL-issued arrest
notices or diffusions (i.e., an NCB-issued arrest request) in 2006 and 2007. Based on past
transnational terrorist events, we then devise a means for translating arrests into fewer incidents
3
and fewer casualties under 12 alternative arrest scenarios. These scenarios allow the reader to
see a range of benefit measures, based on conservative to extremely conservative
counterfactuals. Once we estimate the reduction in terrorist events attributed to INTERPOL’s
coordinating actions, we value these reductions in terms of gross domestic product (GDP)
savings and fewer casualties (i.e., reduced deaths and injuries). The GDP calculations are based
on an econometric model of Blomberg, Hess, and Orphanides (2004) that estimates the impact of
transnational terrorist events on income per capita growth. Finally, we can combine the benefit
and cost information to compute a benefit-cost ratio ( B C ), which indicates the return per dollar
of costs from INTERPOL’s fight against terrorism. We err on the side of caution by
emphasizing costs and downplaying benefits whenever a judgment must be made.
A second purpose here is to use our payback calculations to offer some policy
recommendations with respect to INTERPOL. In particular, our analysis shows that the
cooperative-based counterterrorism activities of INTERPOL have much to offer at a surprisingly
low cost, especially when compared with standard proactive and defensive responses (Sandler,
Arce, and Enders 2009). For example, US homeland security and US proactive responses cost
tens of billions of dollars each in recent years (Treverton et al. 2008). Currently, INTERPOL’s
counterterrorism activities cost just tens of millions of dollars. We show that INTERPOL’s
benefit-cost ratios range from a high of 370 to a low of 65 when GDP savings are included. For
the various arrest scenarios, the ratios have an average value of about 200, thereby returning
$200 for each dollar spent. If, instead, the ratios only include the value of reduced casualties,
then they range from a high of 162 to a low of 4.2, with an average of over 40. Throughout the
study, we raise caveats because we are trying to measure a very elusive entity that has no direct
observation. Despite these caveats, we believe that the payback from further use of INTERPOL
assets is very large. This should not be surprising because network externalities may result in
4
increasing returns to scale as all linked members benefit. Unlike proactive military action by one
country or group of countries, which may breed new grievances and backlash terrorism,
INTERPOL-assisted arrests and vigilance are directed at suspects, while respecting the Universal
Declaration of Human Rights, and are thus much more benign.
The remainder of the paper contains eight sections. Section 2 provides preliminaries,
necessary to understand the role of INTERPOL in fighting transnational terrorism. In Section 3,
we present some theoretical considerations to explain the undersupply of proactive measures, as
well as the underutilization of INTERPOL’s assets. Section 4 identifies the costs of
INTERPOL’s counterterrorism efforts, while Section 5 presents the counterfactual scenarios for
computing the benefits of INTERPOL’s counterterrorism actions. In Section 6, we calculate the
benefits from INTERPOL resources, based on fewer casualties and saved GDP losses associated
with various arrest scenarios. Benefit-cost ratios are then computed in Section 7. Section 8
estimates reduced incidents coming from US instantaneous access at border crossings to
INTERPOL’s SLTD database. These additional benefits underscore the conservative nature of
the benefit-cost calculations in Section 7. Finally, Section 9 contains concluding remarks.
2. INTERPOL and Its Counterterrorism Efforts
To appreciate what INTERPOL contributes to global collective action against transnational
terrorism, the reader must first understand something about the organization and its
counterterrorism assets. The General Assembly governs INTERPOL and consists of member
countries’ delegates, who determine the institutional finances, policies, initiatives, and methods.2
Each member country casts a single (unweighted) vote in the General Assembly, which meets
but once a year. INTERPOL’s main mission is to facilitate as widely as possible transnational
police cooperation in its six priority crime areas, mentioned earlier. The Executive Committee
5
implements the policies of the General Assembly and consists of 13 members, elected by the
General Assembly.
The Secretary General is nominated by the Executive Committee and confirmed by the
General Assembly based on one-country, one-vote secret ballot voting procedure. The Secretary
General directs the General Secretariat and its regional offices in their daily operation in
facilitating better police practices and efficiency worldwide. The General Secretariat and the
regional offices achieve INTERPOL’s mission by interfacing with member countries’ NCBs, the
United Nations, and all organizations, authorities and services, whose mission is to prevent or
combat international crime.
At the outset, we must emphasize that INTERPOL’s General Secretariat Headquarters
and Regional Offices do not arrest criminals or terrorists; rather, they help member countries
make arrests through the use of the organization’s linkages and resources. In the case of
terrorism, INTERPOL assists its member countries in their counterterrorism activities through its
secure communication network (I-24/7), its databases (e.g., SLTD, terrorist profiles, DNA, and
fingerprints), its investigative resources and specialized projects, its money laundering expertise,
its dissemination of best practices (INTERPOL 2008b), and secure communications among
police worldwide. I-24/7 connects all member countries to a secure global communication
system that provides continual access to INTERPOL’s databases. I-24/7 not only allows police
forces to share their data (INTERPOL 2007a), but also gives member countries access to
INTERPOL’s growing dataset on wanted terrorists. This network is used to issue Red Notices
seeking the arrest internationally of suspected criminals, including terrorists. Special
INTERPOL-UN Security Council Notices are issued to prevent international travel, possession
of firearms, and movement of money by suspected members or affiliates of al-Qaida or the
Taliban, while Orange Notices warn of dangerous disguised weapons or acts that present a public
6
hazard (INTERPOL 2008c). A warning of a new terrorist mode of attack could be issued by an
Orange Notice. In addition to Notices, NCBs can use I-24/7 to issue so-called “diffusions,”
which are wanted persons alerts that include information helping to identify the persons.
INTERPOL’s secure communication network is also used to assist in freezing terrorist assets,
issuing travel bans, and preventing arms trafficking. The MIND/FIND internet linkage offers
member countries instant global access to INTERPOL databases that can be applied at border
crossings and elsewhere to catch suspected terrorists. As of November 6, 2009, 46 member
countries, including the United States, have implemented MIND/FIND.
Public Safety and Terrorism (PST) is most responsible for coordinating INTERPOL’s
counterterrorism activities, including running best-practice workshops, the bioterrorism program
(BioT), and INTERPOL Weapons and Explosive Tracking System (IWETS). Both PST and
INTERPOL’s Command and Coordination Center are involved with Incident Response Teams
that are dispatched at countries’ requests to assist in the investigation of terrorist incidents. For
example, teams were dispatched to investigate the Bali nightclub bombings (October 2002), the
Madrid commuter train bombings (March 2004), the Amman Hotels bombings (November 2005)
and the Mumbai attacks (November 2008). For 2002-2008, Incident Response Teams were
dispatched 46 times.3 As part of PST, “the Fusion Task Force (FTF) was created in 2002 to
initiate a proactive, multi-disciplinary approach to assist member countries in terrorism-related
investigations” (INTERPOL 2008b). The FTF works with member countries to collect
information on terrorist groups and their members. This information is accessible to member
countries, unless a country-specific limitation applies, through INTERPOL global police
databases. The FTF focuses on six regions – Southeast Asia, Central Asia, South America,
Africa, the Middle East, and Europe. These task forces also track the movement of suspected
terrorists and post warnings on INTERPOL’s secure website. As a consequence, some Asian-
7
based terrorists have been apprehended in Europe as they flew into an airport hub in Paris.4 FTF
databases are accessed by member countries and have resulted in the arrest of wanted terrorists
and their supporters. These task forces also organize working group meetings to inform law
enforcement agents of regional terrorist risks. In recent years, the FTF has proven to be a
valuable asset in assisting countries to arrest individuals who were either planning or financing
terrorist attacks.5
Another important INTERPOL counterterrorism asset is the Command and Coordination
Center (CCC), which was first started on a temporary basis on 9/11. The CCC is the contact
point for member countries confronting a crisis situation, such as a terrorist attack. CCC analysts
are available 24/7 to answer member countries’ requests during crises and to coordinate the
delivery of special assistance in the four official languages – English, French, Spanish, and
Arabic – of INTERPOL. The CCC also facilitates the exchange of intelligence and the
transmission of all INTERPOL Notices.
In 2003, INTERPOL established the Fugitive Investigative Service (FIS) at the General
Secretariat in Lyon. The FIS offers operational support to member countries in their
international fugitive investigations, including those involving transnational terrorists. This
support includes not only best-practices conferences, but also a network for connecting fugitive
units worldwide. Such linkages facilitate the exchange of information and pooling of expertise.
A final counterterrorism asset is INTERPOL’s Bioterrorism Prevention Unit (BioT)
which augments national and international capabilities to recognize and offset bioterrorism
threats (INTERPOL 2007b). BioT promotes international awareness of bioterrorism through
information exchange, best practices, and workshops. This prevention unit also lobbies for
legislation worldwide that protects against bioterrorism attacks through safeguards and other
means. Funding for BioT has come from the Sloan Foundation, the European Commission, the
8
US State Department, and the Canadian government (INTERPOL 2007b).
In summary, INTERPOL provides a wide range of counterterrorism resources and tools
that can be used by member countries in arresting suspected terrorists and their supporters. Since
INTERPOL does not deploy agents to make the arrests, our measurement of the benefits from
INTERPOL’s counterterrorism activities is necessarily by inference. The most direct measure of
INTERPOL’s success in curbing transnational terrorism is from terrorism-related arrests, where
notices and/or diffusions had been issued. We know that arrests included individuals suspected
of financing, planning, or carrying out terrorist incidents. In only a few instances, do we know
the person arrested – e.g., on 18 August 2005, a suspected bomber in the 3/11 Madrid train
bombing was arrested in Serbia by border authorities who used I-24/7 and INTERPOL’s
databases. A Red Notice had been issued in March 2004 for this suspected terrorist (INTERPOL
2005). In addition, we do not know the final disposition of those arrested – e.g., whether they
were convicted.
3. Theoretical Considerations
There is an important distinction between defensive and protective countermeasures against
transnational terrorism. Defensive actions harden potential targets at home – physical and
human. Such actions dissuade terrorists by reducing their anticipated net gains from attacks.
This follows because defensive measures raise terrorist attack costs and lower their anticipated
benefits. Expected benefits fall because defensive measures reduce terrorists’ logistical success
rates and their gain if successful. A country’s defensive measures have a downside since they
induce the terrorists to find softer targets abroad. The literature shows that defensive measures
against transnational terrorism by one country result in a negative externality to another country
as the attack is displaced abroad (Enders and Sandler 2006a; Sandler and Lapan 1988).
9
Following 9/11, Enders and Sandler (2006b) established that homeland security in North
America and Europe shifted terrorist attacks to the Middle East and Asia where targets were
softer. As countries choose their defensive measures independently, a “defense race” can ensue
with too much being spent as nations compete to transfer attacks abroad (Sandler 2005).
In contrast, proactive or offensive measures directly attack the terrorists, their resource
base (e.g., training camps, finances, or planners), or their supporters. Effective proactive
measures that weaken a transnational terrorist group provide positive externalities to all at-risk
countries. When a country chooses its proactive response to a transnational terrorist threat, the
country equates its perceived marginal benefits to its marginal proactive costs, thereby ignoring
the marginal benefits that its response confers on other countries (Bandyopadhyay and Sandler
2010; Sandler and Siqueira 2006). This then results in too little proactive measures from a social
viewpoint being supplied against the common transnational terrorist threat. As such, the
terrorists profit from the failure of nations to internalize the associated positive externality. In
essence, proactive measures are a pure public good to all potential target countries. Those
countries, whose people or property are most in jeopardy at home or abroad, are most inclined to
take offensive measures, with the other nations free riding on their actions.
There is, thus, a clear role for an international institution, such as INTERPOL, to bolster
these proactive measures by coordinating an international response to transnational terrorism.
Secretary General Ronald K. Noble recognized this need following 9/11 and redirected some of
INTERPOL’s resources into efforts to coordinate counterterrorism actions against al-Qaida and
other transnational terrorist groups (e.g., Jemaah Islamiyah). Moreover, INTERPOL correctly
focused on proactive, rather than defensive, measures since only the former is undersupplied.
We now present a stylized game representation of INTERPOL to explain why some
member countries take advantage of INTERPOL’s resources while others do not. This
10
representation suggests that INTERPOL has reduced the suboptimality associated with proactive
measures but has not eliminated it. To show this outcome, we assume two sets of INTERPOL
members: those that do not perceive a net benefit from participating and those that do perceive a
net benefit from participating. We begin with the former countries which view their benefits, bi,
as less than their costs, ci, from participating. These costs include not only the investment in the
necessary linkages, but also any perceived loss in autonomy from sharing information. A
stylized game is presented in matrix a of Table 1 for six such countries. The number of these
countries is immaterial to our point. Nations that do not view themselves at risk from
transnational terrorism will envision a small bi from utilizing some or all of INTERPOL
proactive assets. With bi small and ci increasing with new technological requirements or
autonomy concerns about information sharing, these countries will likely view their ci as greater
than bi.
In matrix a, we assume six symmetric countries whose 5ib = and whose 8.ic = Each of
these countries has two strategies – not to use INTERPOL’s assets (not be proactive) or to use
INTERPOL’s assets (be proactive). Everything is presented from country i’s viewpoint so that
the payoffs are those of i, under alternative action scenarios of the other five countries. If
country i is not proactive and no other country is proactive, then i obtains 0. If, however, i is not
proactive and one other country is proactive, then country i gains a payoff of 5 in the second cell
on the top row. In the top row, country i gains a benefit of 5 from every proactive country
because INTERPOL’s coordinated proactive response is a pure public good. Thus, the payoffs
in the top row of matrix a increase by 5 with each proactive country.
[Table 1 near here]
In the bottom row of matrix a, country i is proactive and confers a benefit of 5 on itself
and on the other five countries at a cost of 8 to itself. When country i proacts alone, it nets
11
( )3 5 8 .− = − If, however, a second country engages in INTERPOL’s coordinated proactive
measures, then country i receives 2, which is the difference between 10 in benefits (5 from its
actions and 5 from those of the other country) less its costs of 8. The other entries are computed
in a similar fashion. Country i has a dominant strategy not to be proactive, since the payoffs in
the top row are higher than the corresponding payoffs in the bottom row. As all six countries
choose their dominant strategy, the Nash equilibrium results, with no one participating in the
INTERPOL proactive link. The social optimum, however, requires all six countries to
participate for a gain of 22 apiece. The underlying game is a Prisoner’s Dilemma with a
disappointing payoff of 0. Nevertheless, country i will not unilaterally change its strategy at the
Nash equilibrium because 0 3.> −
Next, we focus on those countries that do perceive a net benefit from utilizing
INTERPOL’s proactive assets, despite the associated costs. Without loss of generality, we
assume there to be three (symmetric) countries in this position with, say, 11and 8i ib c= = ; any
( ) 0i ib c− > will do. The benefits are high because these countries are the target of more
transnational terrorist attacks at home or abroad, as in the case of the United States,6 the United
Kingdom, and France. Given the public nature of proactive measures, every proactive country
again confers a benefit of 11 on the countries that view terrorist risks as they do. In the top row
of matrix b, country i receives 0 when no one else utilizes INTERPOL’s assets. Country i
receives 11 or 22 depending on whether one or two countries engage in INTERPOL-coordinated
proactive measures. In the bottom row, country i nets 3 when it acts alone, while it nets another
11 for each country that joins its proactive efforts. The dominant strategy for this three-country
game is for nation i and, therefore, the other two nations to jointly act, thereby giving each of the
three nations a gain of 25 apiece.
12
We now examine the game from the collective viewpoint of all nine countries where
5ib = for the group of six and 11ib = for the group of three. Moreover, all nine countries face
8.ic = The proactive response from the group of three gives a free-rider benefit of ( )15 3 5= ×
to the group of six. In consequence, a constant of 15 is added to the payoffs in all 12 cells in
matrix a. The dominant strategy in the transformed matrix a (not shown) is not to participate,
insofar as each payoff in the top row remains higher by 3 than the corresponding payoff in the
bottom row. As a consequence, these countries gain just 15 and confer no benefits on the group
of three, whose payoff matrix remains unchanged from that in matrix b.
Thus, it is likely to have some member countries of INTERPOL not taking advantage of
the counterterrorism assets that it provided after 9/11. Currently, only 46 members utilize the
instantaneous access to INTERPOL databases offered by MIND/FIND technology owing to high
costs that likely involve autonomy worries. Nonetheless, INTERPOL has greatly improved
social welfare because of the countries that use its resources, without coming near to achieving a
social optimum. These gains in social welfare are apt to produce very large benefit-cost ratios,
given the public nature of the benefits and the relatively small costs involved.
There are a few factors that will raise the share of INTERPOL member countries utilizing
its proactive assets over time. First, as the prime-target countries harden their targets at home,
terrorist attacks will shift abroad. This shift will make these soft-target countries come to value
proactive measures greater, thereby raising bi relative to ci. Second, prime-target countries may
wish to subsidize countries where they have vulnerable interests, thereby lowering these
countries’ cis. Given that INTERPOL offers a network externality, rich target countries have
much to gain from subsidizing other countries’ proactive participation, which weakens the
common terrorist threat. In the stylized game in Table 1, the three participating countries would
gain 11 from each country in the nonparticipating group that begins to use INTERPOL’s
13
linkages. These potentially large gains provide incentives for the three participating countries to
subsidize the efforts of the nonparticipants, thereby reducing their costs. Third, for some
countries, ci is high because of the value that they place on autonomy over their security. As
these countries come to realize that INTERPOL does not greatly infringe on this autonomy,
perceived ci may fall and participation may increase.
4. Costs of INTERPOL’s Counterterrorism Efforts
Insofar as INTERPOL engages in five priority crime areas other than terrorism, we cannot assign
the entire costs of INTERPOL to counterterrorism activities. The INTERPOL website gives the
overall budget and the assessment percentages, from which we can ascertain the dues of each
member country. The website does not provide the breakdown by activity or sub-directorate, so
that we cannot determine INTERPOL’s counterterrorism spending per se. We, thus, contacted
the General Secretariat in Lyon and were invited to come and discuss our data needs, which we
did in October 2008. INTERPOL’s counterterrorism resources involve the PST sub-directorate,
I-24/7, BioT, Fugitive Investigative Service, CCC, and CCC major event spending. The Fusion
Task Force’s costs and workshops expense are part of PST spending. INTERPOL staff members
sent us Excel sheets with the requested budget information, summarized in Table 2 for 2006,
2007, and 2008. In all cases, we used the actual expenditures rather than the budgeted allocation,
since the budget allowed for some fungibility among categories. All spending values are in
euros, except for the next-to-last row where we used the average exchange ratios ($/euros) for
the relevant years to convert INTERPOL’s counterterrorism expenditures into US dollars. These
totals are approximately 23 percent and 26 percent of INTERPOL’s operating expense for 2006
and 2007, respectively, which seem reasonable given the other crime-fighting activities of
INTERPOL. These total operating expenses are given in the last row of Table 2.
14
[Table 2 near here]
We list the partial counterterrorism costs in 2008 for the reader’s interest. In 2008, the
expenditure on I-24/7 is smaller than in earlier years, because the figure, provided to us, only
includes operating expense and excludes associated IT and other infrastructure cost. We cannot
do our B C estimates for 2008 because we do not have the arrest data for that year.
Some caveats are in order. Although the PST is primarily focused on counterterrorism,
PST serves a few other functions in terms of public safety and police training. Member countries
use both I-24/7 and CCC for things other than counterterrorism. For example, the CCC is also
used to address natural disasters, such as a tsunami or an earthquake. I-24/7 can also facilitate
the arrest of a pedophile or a drug trafficker. Moreover, the Fugitive Investigative Service helps
to capture some nonterrorists. By including their entire costs as part of INTERPOL
counterterrorism spending, we are surely erring on the side of overestimating this spending. This
overestimation is intentional because we want any biases in our study to be working against the
hypothesis that INTERPOL’s counterterrorism B C is large, so that the reader has greater
confidence in the conservative nature of our estimated ratios. There are no finer budget
breakdowns that would allow us to better pinpoint INTERPOL’s counterterrorism spending.
5. Counterfactual Scenarios
The most challenging calculation for this study is to compute INTERPOL’s counterterrorism
benefits, since it involves a counterfactual: i.e., how many more transnational terrorist incidents
and associated casualties would there have been had INTERPOL coordinating actions not taken
place. This is challenging because this counterfactual is obviously not observable. Thus, we
must use the members’ utilization of INTERPOL’s counterterrorism assets to estimate the fewer
number of transnational terrorist events. Ultimately, all INTERPOL assets, indicated in Sections
15
2 and 4, are used to arrest terrorists. We, therefore, pin our counterfactual exercise on
INTERPOL-assisted arrests coming from INTERPOL-issued notices or diffusions, including
following alerts of escapes. INTERPOL provided us with the arrest data in two forms: all arrests
and terrorist-related arrests. Obviously, only the latter is germane to our study. In Table 3,
INTERPOL-assisted terrorist-related arrests are broken down by region and by notices and
diffusions for 2006 and 2007. In 2006, there were 74 terrorist-related arrests, while in 2007,
there were 104 terrorist-related arrests. Two-thirds of these arrests took place in Europe in 2006,
while just under three-quarters of these arrests took place in Europe in 2007. This unequal
regional distribution is probably due to two factors: European countries’ proclivity for utilizing
INTERPOL linkages, and the presence of major airport hubs in Europe. A terrorist apprehended
in London may be en route to somewhere else and may belong to al-Qaida or some other terrorist
group. The location of the arrest does not necessarily indicate the terrorist’s planned venue of
attack, nor the terrorist’s network affiliation. Unfortunately, we do not have information about
those arrested.
[Table 3 near here]
To lend credence to our counterfactual computation of benefits, we include myriad
scenarios to show the robustness of our B C estimates. The mean of these B C estimates
represent a central tendency. We rely on International Terrorism: Attributes of Terrorist Events
(ITERATE) data to compute our various scenarios (Mickolus et al. 2008). In particular, we use
ITERATE to translate INTERPOL arrest data into fewer transnational terrorist attacks, reduced
injuries, and fewer deaths. We utilize three time periods – 1968–2007, 2000–2007, and 2006–
2007 – as baselines for our scenarios. In Table 4, these three time periods underlie our
calculations of the average number of nonterrorists injured per incident, the average number of
nonterrorist deaths per incident, and the average size of a terrorist attack force.
16
[Table 4 near here]
For 1968–2007, we apply ITERATE data to make these three calculations for all
incidents, for all incidents with no missing values, and for all incidents with 9/11 included. If an
incident does not have figures for casualties, deaths, or the attack force size, then it is excluded
from the second set of calculations. The average attack force size increases greatly when
incidents with missing values are dropped because media accounts of simple incidents – a bomb
in a trash can – often do not record the attack force size, which usually involves just one or two
terrorists. Media coverage of more complex hijackings or armed attacks frequently includes the
size of the attack force, which involves 4 to 6 persons. Hence, an average force of 1.52 terrorists
is too low, while an average force of 6.13 terrorists is too high. In the third row of Table 4, we
include the two hijacked airplanes that crashed into the World Trade Center. The huge number
of injuries and deaths for these two incidents make them obvious outliers. Since the exact
casualties for these two events are not apportioned to the two incidents in ITERATE, we use the
figures given in Wikipedia.7 The other two 9/11 skyjackings are included in all of our
calculations insofar as the exact casualty counts are known and are not large outliers. We make
similar per-incident injuries, deaths, and attack force calculations for 2000–2007 and 2006–2007,
except that 9/11 is not relevant for the latter period. There are more injuries and deaths per
incident for these latter baseline years. This follows because, prior to the mid-80s, the left-wing
terrorists dominated and they did not strive for large casualty counts. The rise of fundamentalist
terrorists and the greater use of suicide missions raise casualty figures – especially injuries – in
recent years (Sandler, Arce, and Enders 2009).
Our working hypothesis is that each INTERPOL-assisted arrest may result in at most one
fewer transnational terrorist incident. Our much more conservative counterfactual hypothesis is
that the average sized attack force must be apprehended to limit incidents by one in a given year.
17
We do not presume a decrease of one incident in perpetuity for the various scenarios, because
terrorist organizations will recruit new operatives. Capturing one or more terrorist cell members
will surely disrupt a planned incident until new terrorists can be trained or recruited for the
mission. Our presumption is that arrests under various scenarios result in less incidents that year,
because it takes time for terrorists to put the mission back on track. Our hypothesized
counterfactuals are then applied to the three baseline periods in Table 4 to produce 12 scenarios
of fewer incidents, injuries, and deaths for 2006 and 2007. Implicitly, we assume that these
arrests would not have been made in that year without the notice or diffusion.8
The previous paragraph raises the worry of an intertemporal substitution, whereby fewer
terrorist events today, owing to INTERPOL-assisted arrests, are merely replaced by terrorist
incidents tomorrow. Hence, we examined the total number of transnational terrorist incidents in
recent years using ITERATE data. There were 83 incidents in 2006, 93 incidents in 2007, and
92 incidents in 2008. This distribution does not suggest an intertemporal substitution.
Moreover, there is less transnational terrorism in 2006 and 2007, compared to 2004 and 2005
when there were 233 and 108 incidents, respectively.
[Table 5 near here]
In Table 5, Scenario 1 for 2006 involves 74 fewer incidents if each arrest gives one fewer
terrorist incident. This then results in ( )161 74 2.17= × fewer injuries and ( )77 74 1.04= ×
fewer deaths, using the ITERATE-based averages from Table 4. Analogous computations are
made for 2007. Scenario 2 follows similarly, except that 9/11 casualties are included so that
injuries fall by ( )196 74 2.65= × and deaths by ( )92 74 1.24= × . Other calculations for
Scenario 2 follow in an identical manner. Scenario 3 requires the removal of 1.52 terrorists for
one fewer incidents, so that incidents dropped by ( )49 74 1.52= in 2006. The averages from
18
the first row of Table 4 are applied for Scenario 3’s calculations in Table 5. Scenario 4
corresponds to no missing value averages and the necessary removal of 6.13 terrorists to reduce
incidents by one, so that ( )12 74 6.13= fewer incidents are hypothesized in 2006. These fewer
incidents are then multiplied by 2.21 and 1.05 – the per-incident averages associated with no
missing values – to give 27 and 13, respectively.
The eight remaining scenarios are computed in the same way for 2006 and 2007. The
somewhat different scenario is Scenario 9, where we include 9/11 and the large 4.76 attack force.
A similar scenario for 1968–2007 would not have added much variation since casualty rates per
incident did not differ much owing to the larger time period that smoothes out differences. There
are two fewer scenarios for 2006–2007 because there is no 9/11 outlier to worry about. All
calculations follow as before and require no further explanation.
Each of the three baselines results in different hypothesized INTERPOL-assisted
reductions in incidents and casualties. Generally, the smallest savings in benefits come from the
1968–2007 calculations, while the greatest come from the 2000–2007 baseline. Also there are
larger derived benefits from the fewer incidents and casualties in 2007, compared with 2006,
because there were more INTERPOL-assisted arrests in 2007.
Before moving to the benefit-cost calculations, we should address a final concern
because, as mentioned earlier, a large share of INTERPOL-assisted arrests occurred in Europe.
In particular, the size of the terrorist attack force may differ between Europe and the rest of the
world, so that our worldwide attack force averages in Table 4 may bias our results to high
benefit-cost ratios. Thus, we used ITERATE data to compute the size of terrorist attack forces in
Europe for various scenarios.9 We discuss only the values for 1968–2007 because the other
periods’ values follow the same tendency. European attack forces averaged 1.01 terrorists for all
incidents and 4.15 terrorists for incidents with no missing values. Using these values, instead of
19
1.52 and 6.13, would, ceteris paribus, increase benefits because fewer terrorists must be arrested
for there to be one fewer incident. Hence, our use of worldwide averages underestimates
benefits and works against high benefit-cost ratios. We prefer the worldwide averages because
we do not know which country the arrested terrorist really planned to attack.
6. Benefit Measures
We must devise benefit measures that can be fitted into the twelve scenarios for each of the two
sample years. There are two basic drivers of benefits from reduced terrorism: (i) the value of
fewer injuries and deaths and (ii) the reduced losses in GDP.
6.1. Valuing Saved Casualties
We derive a proxy for the injury component of terrorist incidents by appealing to a study
conducted by Abenhaim, Dab, and Salmi (1992) of the injury and psychological consequences of
21 terrorist attacks – 20 bombings and 1 machine-gun attack – that occurred in France during
1982–1987. These authors sent surveys to the population of 324 civilian victims, registered by
the police as having been hospitalized. Abenhaim, Dab, and Salmi (1992) reported an overall
participation rate of 86 percent. Psychological symptoms were identified from assessing
respondents’ self-reported symptoms on a diagnostic portion of the survey. We use their survey
data to create a composite distribution of injuries to apply to the number of wounded, given in
Table 5 for the various scenarios. Although transnational terrorism includes more than
bombings, the French survey on assessing injuries is an excellent proxy because over half of all
transnational terrorist attacks are bombings, which are the most likely attacks to result in injuries.
The distribution of physical injuries and psychological trauma found by Abenhaim, Dab,
and Salmi (1992) for a composite terrorism attack is displayed in Table 6. For example, about
20
47 percent of the injured had hearing loss, 15.4 percent had severe burns, and so on. The injury
total is greater than 100 percent because victims may experience multiple wounds. In Table 6,
the disability weight for each type of injury is taken from Mathers, Lopez, and Murray (2003).
Avoided injuries are converted into avoided deaths by summing over the product of the
percentage of each injury type, jI , and the associated disability weight, jD . The latter measures
the relative value of a health state on a [0,1] scale, where 0jD = is full health, 1jD = is death,
and ( )0,1jD ∈ is an injury. For example, a hearing loss has a disability weight of 0.12 and a
head trauma has a disability weight of 0.35.10 The weighted value of incidence of injury is
j jj
I D∑ , which constitutes the injury equivalent of death. If, for example, all injuries were
deaths, then j j jj j
I D I=∑ ∑ for which jI denotes the type of death rather than the type of
injury. Given the number of injuries, N, avoided for a particular scenario, j jj
N I D×∑ is then
the number of fewer injuries avoided, as measured in terms of deaths avoided. In particular,
j jj
I D∑ is approximately 0.57, implying that every injury avoided in a composite terrorist event
translates into just under 0.57 deaths avoided. The total number of “injuries-as-deaths” avoided
for each scenario is multiplied by $2 million to derive the injury benefits reported later in Section
7. We use a value of life of $2 million because that was the average compensation paid to those
who lost a family member in 9/11 (Sandler, Arce, and Enders 2009).
[Table 6 near here]
We note that our injury calculation is biased downward because it only accounts for the
psychological trauma experienced by those directly affected by the attack. There is now
substantial evidence that major terrorist events have significant psychological impact on the
21
general population. For example, Galea et al. (2002) found that, within two months following
9/11, 7.5 percent of respondents in Manhattan reported symptoms consistent with post traumatic
Fugitive Investigative Service 215,772 181,409 183,608b
Command and Coordination Center (CCC) 1,176,123 1,511,513 1,462,773b
CCC Crisis Major Event 0 358 355,402b
Total counterterrorism expense in euros 10,757,282 12,146,189 8,168,233
Total counterterrorism expense in US dollarsc 13,506,843 16,647,566 12,013,837 INTERPOL operating expense in eurosd 46,991,000 47,222,000 54,621,000 aThe I-24/7 figure for 2008 only includes operating expense and does not include IT and other infrastructure expense. These support expenses are included for the external audited numbers for 2006 and 2007. bBased on year-to-date spending through October 2008 which was 83% of anticipated spending. Hence, the cumulative spending in October 2008 was divided by .83 to give this entry. cBased on the average exchange rate for relevant year, using Eurostat rates of 1.2556 $/euro for 2006, 1.3706 $/euro, and 1.4708 $/euro. dSource: INTERPOL (2008d, 2009).
Tab
le 3
. IN
TE
RPO
L te
rror
ist-
rela
ted
arre
st d
ata
for
2006
and
200
7
2006
2007
R
egio
n A
rres
ts/N
otic
es
Arr
ests
/Dif
fusi
ons
A
rres
ts/N
otic
es
Arr
ests
/Dif
fusi
ons
Afr
ica
2
Am
eric
as (
Nor
th &
Sou
th)
3
2
6
Asi
a Pa
cifi
c
3
Eur
ope
10
39
21
55
Nor
th A
fric
a/M
iddl
e E
ast
8
3
1
Unk
now
n 2
10
2
11
Tot
al
15
59
37
67
Sou
rce:
IN
TE
RP
OL
Gen
eral
Sec
reta
riat
Table 4. Number of casualties and number of terrorist operatives per incidents
Time interval Incidents considered
Nonterrorists injured per
incident
Nonterrorists killed per incident
Number of terrorists in attack force
1968–2007 All incidents 2.17 1.04 1.52 Incidents with no missing valuesa 2.21 1.05 6.13 All incidents with 9/11 2.65b 1.24c 2000–2007 All incidents 6.86 3.54 1.54 Incidents with no missing valuesa 7.36 3.69 4.76 All incidents plus 9/11 12.36b 5.92c 2006–2007 All incidents 3.73 2.63 1.63 Incidents with no missing valuesa 4.07 2.79 5.24
aFor the per-incident calculations, we excluded incidents where the relevant casualty count or the number of terrorists is missing. b Includes the estimated 6291 injuries from the four 9/11 hijackings. It makes almost no difference if incidents with missing values are excluded. c Includes 2729 deaths from the World Trade Center, the 52 deaths from United Airlines 93, and the 236 deaths from American Airlines 77. It makes almost no difference if incidents with missing values are excluded.
Tab
le 5
. Sc
enar
ios
for
bene
fit-
cost
cal
cula
tions
bas
ed o
n ar
rest
not
ices
and
dif
fusi
ons
20
06
20
07
Sce
nari
os
Few
er
Inci
dent
s Fe
wer
In
juri
es
Few
er
Dea
ths
Fe
wer
In
cide
nts
Few
er
Inju
ries
Fe
wer
D
eath
s 19
68–2
007
Sc
enar
io 1
(al
l inc
iden
ts)
74
161
77
10
4 2
26
108
Scen
ario
2 (
9/11
cas
ualti
es in
clud
ed)
74
196
92
10
4 2
76
129
Scen
ario
3 (
atta
ck f
orce
of
1.52
nee
ded)
a 49
10
6 5
1
68
150
7
1 Sc
enar
io 4
(at
tack
for
ce o
f 6.
13 n
eede
d)b
12
27
13
1
7
38
18
2000
–200
7
Scen
ario
5 (
all i
ncid
ents
) 74
50
8 26
2
104
713
36
8 Sc
enar
io 6
(9/
11 c
asua
lties
incl
uded
) 74
91
5 43
8
104
1285
61
6 Sc
enar
io 7
(at
tack
for
ce o
f 1.
54 n
eede
d)a
48
329
170
6
8 4
66
244
Scen
ario
8 (
atta
ck f
orce
of
4.76
nee
ded)
b 16
11
8 5
9
22
162
8
1 Sc
enar
io 9
(9/
11 a
nd a
ttack
for
ce o
f 4.
76)b
16
198
95
2
2 2
72
130
2006
–200
7
Scen
ario
10
(all
inci
dent
s)
74
276
195
10
4 3
88
274
Scen
ario
11
(atta
ck f
orce
of
1.63
nee
ded)
a 45
16
8 11
8
64
239
16
8 Sc
enar
io 1
2 (a
ttack
for
ce o
f 5.
24 n
eede
d)b
14
57
39
2
0
81
55
a The
num
ber
of a
rres
ts is
div
ided
by
the
aver
age
per-
inci
dent
num
ber
of te
rror
ists
in th
e at
tack
for
ce, w
ith th
e un
ders
tand
ing
that
this
is th
e nu
mbe
r of
arr
ests
req
uire
d to
red
uce
the
num
ber
of in
cide
nts
by o
ne.
b The
larg
er r
educ
tion
in in
cide
nts
refl
ects
the
larg
er p
er-i
ncid
ent n
umbe
r of
terr
oris
ts w
hen
only
inci
dent
s w
ith n
o m
issi
ng
valu
es f
or th
e at
tack
for
ce s
ize
are
incl
uded
. T
his
is a
ver
y co
nser
vativ
e m
easu
re o
f th
e re
quir
ed n
umbe
r of
terr
oris
ts
enga
ged
per
inci
dent
s, s
ince
a lo
t of
mis
sing
val
ues
are
likel
y to
be
one
to tw
o te
rror
ists
.
Tab
le 6
. D
istr
ibut
ion
of in
juri
es a
nd tr
aum
a in
a c
ompo
site
terr
oris
t atta
ck
P
hysi
cal I
njur
ies
P
sych
olog
ical
Tra
uma
H
eari
ng
Los
s Se
vere
B
urns
H
ead
Tra
uma
Eye
In
jury
R
espi
rato
ry
Impa
irm
ent
Frac
ture
s an
d/
or A
mpu
tatio
ns
PT
SD
c M
ajor
D
epre
ssio
n P
erce
nt o
f In
juri
esa
I
j 46
.9%
15
.4%
15
.0%
13
.0%
67
.0%
45
.0%
15.1
%
13.3
%
Dis
abil
ity
Wei
ghtb
D
j 0.
12
0.26
0.
35
0.11
0.
28
0.35
0.11
0.
76
a Sour
ce: A
benh
aim
, Dab
, and
Sal
mi (
1992
) b So
urce
: Mat
hers
, Lop
ez, a
nd M
urra
y (2
003)
c PT
SD s
tand
s fo
r po
st tr
aum
atic
str
ess
diso
rder
T
able
7.
Los
t GD
P du
e to
Tra
nsna
tiona
l Ter
rori
sm A
ttack
s: S
cena
rios
for
200
6 R
eal G
DP
a
His
tori
cal
Sce
nari
os
Sce
nari
o S
cena
rio
Sce
nari
o S
cena
rios
S
cena
rio
Sce
nari
o (M
illio
ns
Tib
Cos
t 1,
2, 5
, 6, 1
0 3
4 7
8, 9
11
12
C
ount
ry
of U
S $)
Mill
ions
of
Rea
l US
Dol
lars
c A
lban
ia
4,79
9 1
0.43
6 0.
206
0.16
2 0.
055
0.16
0 0.
071
0.15
3 0.
063
Alg
eria
68
,501
3
18.6
82
8.80
6 6.
935
2.36
0 6.
845
3.01
9 6.
568
2.69
6 A
rgen
tina
314,
833
2
57
.242
26
.980
21
.249
7.
231
20.9
74
9.25
1 20
.124
8.
262
Aus
tria
20
9,72
6 1
19.0
66
8.98
7 7.
078
2.40
8 6.
986
3.08
1 6.
703
2.75
2 A
zerb
aija
n 10
,835
1
0.98
5 0.
464
0.36
6 0.
124
0.36
1 0.
159
0.34
6 0.
142
Bah
rain
10
,017
1
0.91
1 0.
429
0.33
8 0.
115
0.33
4 0.
147
0.32
0 0.
131
Ban
glad
esh
61,7
73
1 5.
616
2.64
7 2.
085
0.70
9 2.
058
0.90
8 1.
974
0.81
1 B
elgi
um
251,
495
1 22
.863
10
.776
8.
487
2.88
8 8.
377
3.69
5 8.
038
3.30
0 B
oliv
ia
9,81
8 1
0.89
3 0.
421
0.33
1 0.
113
0.32
7 0.
144
0.31
4 0.
129
Bos
nia/
Her
z.
6,83
6 2
1.24
3 0.
586
0.46
1 0.
157
0.45
5 0.
201
0.43
7 0.
179
Bra
zil
741,
349
1 67
.395
31
.766
25
.018
8.
513
24.6
94
10.8
92
23.6
94
9.72
7 B
urun
di
802
2 0.
146
0.06
9 0.
054
0.01
8 0.
053
0.02
4 0.
051
0.02
1 C
ambo
dia
5,71
3 1
0.51
9 0.
245
0.19
3 0.
066
0.19
0 0.
084
0.18
3 0.
075
Cha
d 2,
570
1 0.
234
0.11
0 0.
087
0.03
0 0.
086
0.03
8 0.
082
0.03
4 C
hile
92
,750
1
8.43
2 3.
974
3.13
0 1.
065
3.09
0 1.
363
2.96
4 1.
217
Chi
na
1,92
8,69
2 1
175.
336
82.6
42
65.0
87
22.1
48
64.2
45
28.3
37
61.6
41
25.3
06
Col
ombi
a 10
0,45
4 2
18.2
64
8.60
9 6.
780
2.30
7 6.
692
2.95
2 6.
421
2.63
6 C
ongo
, Rep
3,
922
3 1.
070
0.50
4 0.
397
0.13
5 0.
392
0.17
3 0.
376
0.15
4 C
ote
d'Iv
orie
10
,350
1
0.94
1 0.
443
0.34
9 0.
119
0.34
5 0.
152
0.33
1 0.
136
Cyp
rus
10,9
44
1 0.
995
0.46
9 0.
369
0.12
6 0.
365
0.16
1 0.
350
0.14
4 C
zech
Rep
68
,569
1
6.23
4 2.
938
2.31
4 0.
787
2.28
4 1.
007
2.19
1 0.
900
Den
mar
k 17
1,22
9 1
15.5
66
7.33
7 5.
778
1.96
6 5.
704
2.51
6 5.
473
2.24
7 E
cuad
or
20,4
03
1 1.
855
0.87
4 0.
689
0.23
4 0.
680
0.30
0 0.
652
0.26
8
Egy
pt,
121,
831
5 55
.378
26
.102
20
.557
6.
995
20.2
91
8.95
0 19
.469
7.
993
Eri
trea
75
1 1
0.06
8 0.
032
0.02
5 0.
009
0.02
5 0.
011
0.02
4 0.
010
Eth
iopi
a 11
,223
1
1.02
0 0.
481
0.37
9 0.
129
0.37
4 0.
165
0.35
9 0.
147
Fran
ce
1,44
1,08
0 3
393.
022
185.
246
145.
894
49.6
45
144.
008
63.5
19
138.
172
56.7
25
Geo
rgia
4,
440
2 0.
807
0.38
1 0.
300
0.10
2 0.
296
0.13
0 0.
284
0.11
7 G
erm
any
1,98
1,75
0 3
540.
477
254.
747
200.
632
68.2
71
198.
038
87.3
50
190.
012
78.0
07
Gre
ece
178,
734
3 48
.746
22
.976
18
.095
6.
157
17.8
61
7.87
8 17
.137
7.
035
Gua
tem
ala
22,7
25
1 2.
066
0.97
4 0.
767
0.26
1 0.
757
0.33
4 0.
726
0.29
8 H
aiti
3,80
4 3
1.03
7 0.
489
0.38
5 0.
131
0.38
0 0.
168
0.36
5 0.
150
Indi
a 65
0,94
0 4
236.
705
111.
568
87.8
68
29.9
00
86.7
32
38.2
55
83.2
17
34.1
64
Indo
nesi
a 20
8,88
2 3
56.9
68
26.8
51
21.1
47
7.19
6 20
.874
9.
207
20.0
28
8.22
2 Ir
an
134,
639
2 24
.480
11
.538
9.
087
3.09
2 8.
970
3.95
6 8.
606
3.53
3 Ir
aq
18,1
62
5 8.
255
3.89
1 3.
065
1.04
3 3.
025
1.33
4 2.
902
1.19
2 Is
rael
13
4,74
5 4
48.9
98
23.0
95
18.1
89
6.18
9 17
.954
7.
919
17.2
26
7.07
2 It
aly
1,14
4,65
5 2
208.
119
98.0
94
77.2
56
26.2
89
76.2
57
33.6
35
73.1
67
30.0
38
Jord
an
11,3
76
4 4.
137
1.95
0 1.
536
0.52
3 1.
516
0.66
9 1.
454
0.59
7 K
enya
15
,266
1
1.38
8 0.
654
0.51
5 0.
175
0.50
9 0.
224
0.48
8 0.
200
Kuw
ait
50,4
50
2 9.
173
4.32
3 3.
405
1.15
9 3.
361
1.48
2 3.
225
1.32
4 K
yrgy
zsta
n 1,
672
1 0.
152
0.07
2 0.
056
0.01
9 0.
056
0.02
5 0.
053
0.02
2 L
eban
on
20,2
43
3 5.
521
2.60
2 2.
049
0.69
7 2.
023
0.89
2 1.
941
0.79
7 L
ibya
40
,508
2
7.36
5 3.
471
2.73
4 0.
930
2.69
9 1.
190
2.58
9 1.
063
Mad
agas
car
4,36
5 1
0.39
7 0.
187
0.14
7 0.
050
0.14
5 0.
064
0.14
0 0.
057
Mal
aysi
a 11
2,34
1 2
20.4
26
9.62
7 7.
582
2.58
0 7.
484
3.30
1 7.
181
2.94
8 M
aldi
ves
952
1 0.
087
0.04
1 0.
032
0.01
1 0.
032
0.01
4 0.
030
0.01
2 M
auri
tani
a 1,
345
2 0.
245
0.11
5 0.
091
0.03
1 0.
090
0.04
0 0.
086
0.03
5 M
exic
o 64
0,04
2 1
58.1
86
27.4
25
21.5
99
7.35
0 21
.320
9.
404
20.4
56
8.39
8 M
oroc
co
47,9
24
2 8.
714
4.10
7 3.
235
1.10
1 3.
193
1.40
8 3.
063
1.25
8 N
ethe
rlan
ds
411,
545
1 37
.413
17
.634
13
.888
4.
726
13.7
09
6.04
7 13
.153
5.
400
Nig
eria
61
,874
3
16.8
75
7.95
4 6.
264
2.13
2 6.
183
2.72
7 5.
932
2.43
6 N
orw
ay
187,
900
1 17
.082
8.
051
6.34
1 2.
158
6.25
9 2.
761
6.00
5 2.
465
Paki
stan
94
,305
5
42.8
66
20.2
04
15.9
12
5.41
5 15
.707
6.
928
15.0
70
6.18
7 Pe
ru
66,4
60
1 6.
042
2.84
8 2.
243
0.76
3 2.
214
0.97
6 2.
124
0.87
2 Ph
ilipp
ines
95
,130
2
17.2
96
8.15
2 6.
421
2.18
5 6.
338
2.79
5 6.
081
2.49
6 R
ussi
an F
ed
353,
493
5 16
0.67
9 75
.734
59
.646
20
.296
58
.875
25
.968
56
.489
23
.191
Sa
udi A
rabi
a 22
4,47
1 3
61.2
19
28.8
55
22.7
25
7.73
3 22
.432
9.
894
21.5
22
8.83
6 So
lom
on
324
1 0.
029
0.01
4 0.
011
0.00
4 0.
011
0.00
5 0.
010
0.00
4 Sp
ain
683,
282
3 18
6.35
0 87
.834
69
.175
23
.539
68
.281
30
.117
65
.514
26
.896
Sr
i Lan
ka
20,0
78
4 7.
301
3.44
1 2.
710
0.92
2 2.
675
1.18
0 2.
567
1.05
4 Su
dan
17,1
43
3 4.
675
2.20
4 1.
736
0.59
1 1.
713
0.75
6 1.
644
0.67
5 S
yria
23
,913
2
4.34
8 2.
049
1.61
4 0.
549
1.59
3 0.
703
1.52
9 0.
628
Tha
iland
15
7,53
1 3
42.9
63
20.2
50
15.9
48
5.42
7 15
.742
6.
944
15.1
04
6.20
1 T
urke
y 33
0,60
2 5
150.
274
70.8
30
55.7
83
18.9
82
55.0
62
24.2
87
52.8
31
21.6
89
Uga
nda
7,78
8 1
0.70
8 0.
334
0.26
3 0.
089
0.25
9 0.
114
0.24
9 0.
102
UK
1,
633,
724
4 59
4.08
1 28
0.01
3 22
0.53
0 75
.042
21
7.67
9 96
.013
20
8.85
7 85
.744
U
S 10
,955
,120
3
2987
.760
1,
408.
244
1,10
9.09
3 37
7.40
1 1,
094.
752
482.
870
1,05
0.38
4 43
1.22
3 U
zbek
ista
n 18
,086
1
1.64
4 0.
775
0.61
0 0.
208
0.60
2 0.
266
0.57
8 0.
237
Ven
ezue
la
132,
155
2 24
.028
11
.325
8.
920
3.03
5 8.
804
3.88
3 8.
447
3.46
8 Y
emen
11
,518
4
4.18
8 1.
974
1.55
5 0.
529
1.53
5 0.
677
1.47
2 0.
605
TO
TA
LS
$6,5
34.6
8 $3
,080
.040
$2
,425
.752
$8
25.4
33
$2,3
94.3
86
$1,0
56.1
10
$2,2
97.3
48
$943
.150
a R
eal G
DP
equa
ls f
ive
year
ave
rage
of
GD
P, v
alue
d in
mill
ions
of
year
200
0 U
S do
llars
. IT
ER
AT
E li
sts
inci
dent
s fo
r C
uba,
the
Gaz
a St
rip,
Q
atar
, Som
alia
, and
Yug
osla
via.
Sin
ce W
orld
Ban
k re
al G
DP
data
is n
ot a
vaila
ble
for
thes
e ar
eas,
we
have
exc
lude
d th
em f
rom
the
calc
ulat
ions
. Fi
gure
s fo
r B
ahra
in, I
raq,
and
Kuw
ait a
re th
e th
ree-
year
ave
rage
s of
200
3-20
05.
b T i e
qual
s th
e nu
mbe
r of
yea
rs w
ith a
t lea
st o
ne te
rror
ist i
ncid
ent.
c A
ll sc
enar
ios
are
valu
ed in
mill
ions
of
year
200
0 U
S do
llars
.
Tab
le 8
. L
ost G
DP
due
to T
rans
natio
nal T
erro
rism
Atta
cks:
Sce
nari
os f
or 2
007
Rea
l GD
Pa
H
isto
rica
l S
cena
rios
S
cena
rio
Sce
nari
o S
cena
rios
S
cena
rio
Sce
nari
o (M
illio
ns
Tib
Cos
t 1,
2, 5
, 6, 1
0 3,
7
4 8,
9
11
12
Cou
ntry
of
US
$)
M
illio
ns o
f R
eal U
S D
olla
rsc
Alb
ania
4,
799
1 0.
436
0.22
9 0.
183
0.06
7 0.
083
0.17
7 0.
077
Alg
eria
68
,501
3
18.6
82
9.81
3 7.
842
2.86
1 3.
543
7.56
7 3.
278
Arg
entin
a 31
4,83
3 2
57.2
42
30.0
67
24.0
28
8.76
7 10
.856
23
.187
10
.043
A
ustr
ia
209,
726
1 19
.066
10
.014
8.
003
2.92
0 3.
616
7.72
3 3.
345
Aze
rbai
jan
10,8
35
1 0.
985
0.51
7 0.
413
0.15
1 0.
187
0.39
9 0.
173
Bah
rain
10
,017
1
0.91
1 0.
478
0.38
2 0.
139
0.17
3 0.
369
0.16
0 B
angl
ades
h 61
,773
1
5.61
6 2.
950
2.35
7 0.
860
1.06
5 2.
275
0.98
5 B
elgi
um
251,
495
1 22
.863
12
.009
9.
597
3.50
2 4.
336
9.26
1 4.
011
Bol
ivia
9,
818
1 0.
893
0.46
9 0.
375
0.13
7 0.
169
0.36
2 0.
157
Bos
nia/
Her
z.
6,83
6 2
1.24
3 0.
653
0.52
2 0.
190
0.23
6 0.
503
0.21
8 B
razi
l 74
1,34
9 1
67.3
95
35.4
00
28.2
89
10.3
22
12.7
82
27.2
99
11.8
24
Bur
undi
80
2 2
0.14
6 0.
077
0.06
1 0.
022
0.02
8 0.
059
0.02
6 C
ambo
dia
5,71
3 1
0.51
9 0.
273
0.21
8 0.
080
0.09
8 0.
210
0.09
1 C
had
2,57
0 1
0.23
4 0.
123
0.09
8 0.
036
0.04
4 0.
095
0.04
1 C
hile
92
,750
1
8.43
2 4.
429
3.53
9 1.
291
1.59
9 3.
415
1.47
9 C
hina
1,
928,
692
1 17
5.33
6 92
.095
73
.598
26
.853
33
.253
71
.022
30
.761
C
olom
bia
100,
454
2 18
.264
9.
593
7.66
7 2.
797
3.46
4 7.
398
3.20
4 C
ongo
, Rep
3,
922
3 1.
070
0.56
2 0.
449
0.16
4 0.
203
0.43
3 0.
188
Cot
e d'
Ivor
ie
10,3
50
1 0.
941
0.49
4 0.
395
0.14
4 0.
178
0.38
1 0.
165
Cyp
rus
10,9
44
1 0.
995
0.52
3 0.
418
0.15
2 0.
189
0.40
3 0.
175
Cze
ch R
ep
68,5
69
1 6.
234
3.27
4 2.
617
0.95
5 1.
182
2.52
5 1.
094
Den
mar
k 17
1,22
9 1
15.5
66
8.17
6 6.
534
2.38
4 2.
952
6.30
5 2.
731
Ecu
ador
20
,403
1
1.85
5 0.
974
0.77
9 0.
284
0.35
2 0.
751
0.32
5
Egy
pt,
121,
831
5 55
.378
29
.087
23
.245
8.
481
10.5
03
22.4
32
9.71
5 E
ritr
ea
751
1 0.
068
0.03
6 0.
029
0.01
0 0.
013
0.02
8 0.
012
Eth
iopi
a 11
,223
1
1.02
0 0.
536
0.42
8 0.
156
0.19
3 0.
413
0.17
9 Fr
ance
1,
441,
080
3 39
3.02
2 20
6.43
6 16
4.97
2 60
.193
74
.539
15
9.19
9 68
.951
G
eorg
ia
4,44
0 2
0.80
7 0.
424
0.33
9 0.
124
0.15
3 0.
327
0.14
2 G
erm
any
1,98
1,75
0 3
540.
477
283.
887
226.
867
82.7
76
102.
504
218.
928
94.8
21
Gre
ece
178,
734
3 48
.746
25
.604
20
.461
7.
466
9.24
5 19
.745
8.
552
Gua
tem
ala
22,7
25
1 2.
066
1.08
5 0.
867
0.31
6 0.
392
0.83
7 0.
362
Hai
ti 3,
804
3 1.
037
0.54
5 0.
435
0.15
9 0.
197
0.42
0 0.
182
Indi
a 65
0,94
0 4
236.
705
124.
330
99.3
58
36.2
52
44.8
92
95.8
81
41.5
27
Indo
nesi
a 20
8,88
2 3
56.9
68
29.9
22
23.9
12
8.72
5 10
.804
23
.076
9.
994
Iran
13
4,63
9 2
24.4
80
12.8
58
10.2
75
3.74
9 4.
643
9.91
6 4.
295
Iraq
18
,162
5
8.25
5 4.
336
3.46
5 1.
264
1.56
6 3.
344
1.44
8 Is
rael
13
4,74
5 4
48.9
98
25.7
36
20.5
67
7.50
4 9.
293
19.8
47
8.59
6 It
aly
1,14
4,65
5 2
208.
119
109.
315
87.3
59
31.8
74
39.4
71
84.3
01
36.5
12
Jord
an
11,3
76
4 4.
137
2.17
3 1.
736
0.63
4 0.
785
1.67
6 0.
726
Ken
ya
15,2
66
1 1.
388
0.72
9 0.
583
0.21
3 0.
263
0.56
2 0.
243
Kuw
ait
50,4
50
2 9.
173
4.81
8 3.
850
1.40
5 1.
740
3.71
6 1.
609
Kyr
gyzs
tan
1,67
2 1
0.15
2 0.
080
0.06
4 0.
023
0.02
9 0.
062
0.02
7 L
eban
on
20,2
43
3 5.
521
2.90
0 2.
317
0.84
6 1.
047
2.23
6 0.
969
Lib
ya
40,5
08
2 7.
365
3.86
9 3.
092
1.12
8 1.
397
2.98
3 1.
292
Mad
agas
car
4,36
5 1
0.39
7 0.
208
0.16
7 0.
061
0.07
5 0.
161
0.07
0 M
alay
sia
112,
341
2 20
.426
10
.729
8.
574
3.12
8 3.
874
8.27
4 3.
583
Mal
dive
s 95
2 1
0.08
7 0.
045
0.03
6 0.
013
0.01
6 0.
035
0.01
5 M
auri
tani
a 1,
345
2 0.
245
0.12
8 0.
103
0.03
7 0.
046
0.09
9 0.
043
Mex
ico
640,
042
1 58
.186
30
.562
24
.424
8.
911
11.0
35
23.5
69
10.2
08
Mor
occo
47
,924
2
8.71
4 4.
577
3.65
8 1.
335
1.65
3 3.
530
1.52
9 N
ethe
rlan
ds
411,
545
1 37
.413
19
.651
15
.704
5.
730
7.09
6 15
.155
6.
564
Nig
eria
61
,874
3
16.8
75
8.86
3 7.
083
2.58
4 3.
200
6.83
5 2.
960
Nor
way
18
7,90
0 1
17.0
82
8.97
2 7.
170
2.61
6 3.
240
6.91
9 2.
997
Paki
stan
94
,305
5
42.8
66
22.5
15
17.9
93
6.56
5 8.
130
17.3
63
7.52
0 Pe
ru
66,4
60
1 6.
042
3.17
3 2.
536
0.92
5 1.
146
2.44
7 1.
060
Phili
ppin
es
95,1
30
2 17
.296
9.
085
7.26
0 2.
649
3.28
0 7.
006
3.03
4 R
ussi
an F
ed
353,
493
5 16
0.67
9 84
.397
67
.445
24
.608
30
.474
65
.085
28
.189
Sa
udi A
rabi
a 22
4,47
1 3
61.2
19
32.1
56
25.6
97
9.37
6 11
.611
24
.798
10
.740
So
lom
on
324
1 0.
029
0.01
5 0.
012
0.00
5 0.
006
0.01
2 0.
005
Spai
n 68
3,28
2 3
186.
350
97.8
81
78.2
21
28.5
40
35.3
42
75.4
83
32.6
93
Sri L
anka
20
,078
4
7.30
1 3.
835
3.06
5 1.
118
1.38
5 2.
957
1.28
1 Su
dan
17,1
43
3 4.
675
2.45
6 1.
963
0.71
6 0.
887
1.89
4 0.
820
Syr
ia
23,9
13
2 4.
348
2.28
4 1.
825
0.66
6 0.
825
1.76
1 0.
763
Tha
iland
15
7,53
1 3
42.9
63
22.5
66
18.0
34
6.58
0 8.
148
17.4
03
7.53
7 T
urke
y 33
0,60
2 5
150.
274
78.9
32
63.0
78
23.0
15
28.5
00
60.8
70
26.3
64
Uga
nda
7,78
8 1
0.70
8 0.
372
0.29
7 0.
108
0.13
4 0.
287
0.12
4 U
K
1,63
3,72
4 4
594.
081
312.
043
249.
367
90.9
85
112.
671
240.
641
104.
225
US
10,9
55,1
20
3 29
87.7
60
1,56
9.32
8 1,
254.
121
457.
585
566.
644
1,21
0.23
2 52
4.16
8 U
zbek
ista
n 18
,086
1
1.64
4 0.
864
0.69
0 0.
252
0.31
2 0.
666
0.28
8 V
enez
uela
13
2,15
5 2
24.0
28
12.6
21
10.0
86
3.68
0 4.
557
9.73
3 4.
215
Yem
en
11,5
18
4 4.
188
2.20
0 1.
758
0.64
1 0.
794
1.69
7 0.
735
TO
TA
LS
$6,5
34.6
8 $3
,432
.357
$2
,742
.952
$1
,000
.807
$1
,239
.336
$2
,646
.959
$1
,146
.435
a R
eal G
DP
equa
ls f
ive
year
ave
rage
of
GD
P, v
alue
d in
mill
ions
of
year
200
0 U
S do
llars
. IT
ER
AT
E li
sts
inci
dent
s fo
r C
uba,
the
Gaz
a St
rip,
Qat
ar, S
omal
ia, a
nd Y
ugos
lavi
a. S
ince
Wor
ld B
ank
real
GD
P da
ta is
not
ava
ilabl
e fo
r th
ese
area
s, w
e ha
ve e
xclu
ded
them
fro
m
the
calc
ulat
ions
. Fi
gure
s fo
r B
ahra
in, I
raq,
and
Kuw
ait a
re th
e th
ree-
year
ave
rage
s of
200
3-20
05.
b T i e
qual
s th
e nu
mbe
r of
yea
rs w
ith a
t lea
st o
ne te
rror
ist i
ncid
ent.
c A
ll sc
enar
ios
are
valu
ed in
mill
ions
of
year
200
0 U
S do
llars
.
Table 9. Benefit calculations based on arrest notices and diffusions (in millions of US dollars) 2006 2007