The Threat of Terrorism in Southern Europe: A Probabilistic Risk Index and Analysis

Table of Contents

§ Executive Summary………………………………………………………………….2

§ Statement of Purpose………………………………………………………………...3

§ Objectives……………………………………………………………….……………

§ Methodology.…………………………………………………………………………4

§ Discussion of Variables……………………………………………………………...5

o Group Variables……………………………………………………………….

§ Size ……………………………………………………………………

§ Funding …………………………………………………………….…6

§ Length of Activity…………………………………………………..…

§ Number of Attacks………………………………………………….…7

§ Date of Last Attack……………………………………………………

§ Typical Target and Method of Attack ………………………………...8

§ Al Qaeda Presence……………………………………………………

o Government Variables ………………………………………………………..9

§ Strength and Technique of Response…………………………………

§ Antagonistic or Agonistic Stance……………………………………..

§ Allegations of Torture or Illegal Imprisonment………………………

§ Rumored Ties to Group (Funding or Harbor)………………………..10

§ Regime Instability…………………………………………………….

§ Corruption and Enforcement of Laws ……………………………….11

o Infrastructure Variables…………………………………………………...….

§ Size, Strength, and Security of Borders………………………………

§ Size, Strength, and Security of Railways, Airports, and Ports………12

o Society (Internal) Variables …………………………………………………13

§ Public Opinion of Group or Government…………………………….

§ Press Opinion of Group or Government…………………………..…14

§ Islamic or Religious Fundamentalism.……………………………….

§ Anti-Americanism (Group).………………………………………….

§ Youth Bulge………………………………………………………….15

o Surroundings Variables.……………………………………………………...16

§ Bad neighborhood…………………………………………………….

§ Weak or Few Political Allies……………………………………...….

§ Member of Relevant International Groups or Treaties……………….

§ Globalization Ranking………………………………………….……17

§ Results / Recommendations….…………………………………………………….

§ Limitations / Future Incarnations of the Model…………………………………..21

List of Tables

§ Table I: Breakdown of Variables …………………………………...………………...5

§ Table II: Breakdown of Group Score Tabulations ……………………...…………….8

§ Table III: Breakdown of Government Score Tabulations...…………………..……...11

§ Table IV: Breakdown of Infrastructure Score Tabulations.….…………………..…..13

§ Table V: Breakdown of Society (Internal) Score Tabulations..….…………………..15

§ Table VI: Breakdown of Surroundings Score Tabulations..…………………………17

§ Table VII: Breakdown of Final Score Tabulations.…….………………………..…..21

Executive Summary

Spain was deemed to be the country most at risk for a terrorist attack. Its high Group scores -- due to the size and activity of its two active terrorist groups -- along with Spain’s large and vulnerable system of state infrastructure make the country particularly susceptible to attack. Greece also looks to be at a heightened risk owing to its high Government and Surroundings scores with their instability and corruption ratings, in addition to the three terrorist groups operating within its borders. The remainder of those countries surveyed seem to be at a much more diminished risk of attack, all of them scoring in the 30s or below. Portugal, somewhat incredibly, scored almost no risk of being attacked, holding the lowest score in four of the model’s five categories.

Statement of Problem

As members of the State Department our section’s goals and overriding interests are firmly entrenched in the war on terror. As such we have been tasked with evaluating the viability of several countries in Europe as sites for future terrorist attacks. Previous reports generated by our department have focused on the likelihood of attacks specifically against American embassies in various countries.

Rather than repeat this research, we have decided to broaden our investigation and point it at the societal level in order to examine the probability of a generalized attack – be it on US interests or those of the host country – in the interest of finding where the weak spots lie in our endeavor to create a safer global society.

Using our model we hope to discover which of the six European countries examined will be most likely to suffer a terrorist attack in the next two to three years. We have focused on the four main countries of southern Europe – Portugal, Spain, Italy, and Greece – in the effort to discern whether the common knowledge regarding their status as “laggards” relative to the rest of Europe hold true. In order to better visualize this contrast we have also included France and Germany in our evaluation to see whether a higher state of industrialization and apparent progress results in a safer society or not.

Objectives

By examining data in a host of categories, encompassing over twenty different indicators under five umbrella variables, we will determine the probability of a terrorist attack in the homelands of our six European allies, find out which is most at risk, and highlight potential problem areas those governments can use to shore up their defenses.

Methodology

As mentioned before, our model evaluated these six countries in a series of five composite categories gauging their risk for future terrorist attack – a terrorist/guerrilla group score, a government score, a state infrastructure score, an internal society score, and one for the country’s surroundings. Beneath these five umbrellas lay four to eight smaller variables that help form each composite score, with higher scores indicating a higher degree of risk.

The overall hypothesis was that the higher the scores a country had in each of the composite variables, the more at risk they would be for future attack. A high Group score would indicate the presence of a large, well funded, well established, and active terrorist/insurgent group in a particular country. A high Government tabulation would mean there is a weak, unstable, antagonistic, and illegitimate ruling coalition in power. A high Infrastructure score would indicate the country is awash in weakly protected, highly abundant possible terrorist targets. A high Society ranking would mean there is a highly volatile, antagonistic situation inside the country’s borders, while a high Surroundings score would mean there is an equally dangerous environment outside of them.

Our model operated on a 100-point system where the closer a country came to 100, the higher at risk they were for a terrorist attack, and each of the five categories (and the underlying variables, as well) was weighted differently according to their perceived importance to the problem. Since this is primarily a showdown between two main combatants – a terrorist/guerrilla group and the state – those categories have more variables forming their composite scores and are weighted accordingly overall. (25 points per category, five points per variable within them.) The other influences cannot be discounted, though, and are thus divided among the remaining points, with slightly more points going towards the Infrastructure variable since it addresses more variables than the others. (20 points spread among four variables; the Society and Surroundings composites received 15 points a piece for the three variables in each.)

Table I: Breakdown of Variables

Group (25)	Government (25)	Infrastructure (20)	Society (15)	Surroundings (15)
Size (2.5)	Strength / Technique of Response (2.5)	Size/ Strength/ Security of Borders (5)	Public Opinion of Group / Government (2.5)	Bad Neighborhood (5)
Funding (2.5)	Antagonistic / Agonistic Stance (2.5)	Size/ Strength/ Security of Railways (5)	Press Opinion of Group / Government (2.5)	Member of Relevant International Groups / Treaties (5)
Length of Activity (5)	Allegations of Torture / Imprisonment (5)	Size/ Strength/ Security of Airports (5)	Islamic / Religious Fundamentalism (2.5)	Globalization Ranking (5)
Number of Attacks (2.5)	Rumored Ties to Group (Funding, Harbor) (5)	Size/ Strength/ Security of Ports (5)	Anti-Americanism (2.5)
Date of Last Attack (2.5)	Regime Instability (5)		Youth Bulge (5)
Typical Target, Method of Attack (5)	Corruption / Enforcement of Laws (5)
Al Qaeda Presence (5)

Discussion of Variables

Group variables:

Size (2.5) – The number of members in a group was considered an important indicator of their overall strength and their ability to attack the state, thus larger groups received higher scores. The breakdown of the ratings and data went along the lines of small, medium, and large – groups with less than 25 members received 1 point, 25-100 received 3, and over 100 received 5 points, the highest risk. Data was culled primarily from the Council on Foreign Relations website,[1] the CRS Report for Congress,[2] and various State Department websites and reports, including several incarnations of their annual Patterns of Global Terrorism report.[3] These sources were used to evaluate all subsequent variables in the Group category; other additional sources are noted within each description.

Funding (2.5) – The sources of funding for a group and its estimated net reserves, if available, were another indicator of the group’s strength, the thinking being that the more funds at a group’s disposal, the greater their stability, probable longevity, and range of options available to them for attack. If a group was constantly struggling to fund its operations, its cohesion and activity would likely be impeded, thus the more they had, the more potent a threat they were, and the higher the score they received. Groups that were not self-sufficient, frequently hampered by their lack of capital, received 1 point (low risk), those that are self-sustaining with occasional surpluses of cash received 3, and those that were well-funded, self-sustaining, and constantly possessing surpluses received 5, the highest risk. Data for this variable was hard to come by and obviously rather subjective since these groups’ finances are secret and highly fluid, but the aforementioned sources in addition to the 2004 CRS report to Congress[4] helped develop a good idea of a group’s fiscal strength. [This variable’s results were averaged with the former’s in an effort to conserve space and due to correlation amongst the pair.]

Length of Activity (5) – The longer a group has been active, the more entrenched they are in society and the harder they will be to control or remove for the state. Thus the more years of activity a group has, the higher their score in the model. Groups that have been active for less than five years posed the lowest risk and received 1 point, while those that are between 5 and 10 years old received 3 points. Groups over ten years old have proven the most difficult to remove and thus pose the highest risk, receiving 5 points.

Number of Attacks (2.5) – Correspondingly, the more active a group is over those years, the more of a threat they pose to the state, and the higher the score they received. Groups that have conducted less than 25 attacks during their existence and caused a comparable amount of casualties received 1 point, groups responsible for 25-100 attacks and similar casualties received 3, while those responsible for over 100 attacks and an parallel number of casualties received 5 points, the highest risk. The International Policy Institute for Counter-Terrorism[5] was consulted in addition to the aforementioned to rank this variable.

Date of Last Attack (2.5) – Scores for this variable reflect how long it has been since a group has been heard from. A long period of dormancy (over 5 years inactive) could reflect several things and was therefore difficult to score -- the effectiveness of the government in eradicating the threat, the group’s dissolution, or their period of silence and planning before striking again. It was difficult to tell which of the three was in play here, but taken in concert with the other variables, we decided that the longer a group was silent, the more unlikely they were to strike again and the lower their score. Cases where groups went from being highly active to long periods of silence were easier to score and deemed an indication of an effective government counterterrorism program, thus their lower scores. Groups that were still active (less than one year inactivity) were deemed the highest threat and received 5 points in the model, while those inactive for 1-5 years were classified as medium threats, receiving 3 points. (The first example of over 5 years inactivity received only 1 point.) [This variable’s results were averaged with the former’s in an effort to conserve space and due to correlation amongst the pair.]

Typical Target and Method of Attack (5) – This variable looks at the methods typically employed by a group – assassinations, bombings, kidnappings – and who they typically target – buildings, leaders, civilians. The more violent a group was and the more they directed their actions at people, be it political leaders or innocent civilians, the higher the score they received. (Those who balanced their attack between the two demographics received 3 points while those whose civilian casualties outnumbered all other parties received 5 points.) If a group was primarily intent upon property damage and spectacular displays, targeting buildings and symbolic entities rather than humans, they received lower scores. (1 point)

Al Qaeda Presence (5) – If a group has no homegrown terrorist group operating within their borders, the last possible threat comes from an al Qaeda cell that may be present. Those countries without a recent al Qaeda presence received 1 point, those that had a cell recently removed (in the three years since 9/11) received 3, and those who still had an al Qaeda presence active and acknowledged received 5 points, the highest risk.

Table II: Breakdown of Group Score Tabulations

Country	Size	Funding	Length of Activity	Number of Attacks	Date of Last Attack	Typical Target, Method of Attack	Al Qaeda Presence	Total
Portugal	0	0	0	0	0	0	0	0
Spain (ETA/GRAPO)	5/1	5/1	5/5	5/3	3/1	5/1	3	22/12
Italy	0	0	0	0	0	0	3	3
Greece (Nov 17/ELA/RN)	3/0/3	3/0/3	5/5/3	3/3/1	3/1/3	3/1/1	0	14/8/9
France (Alone/With ETA)	0/5	0/5	0/5	0/5	0/3	0/5	3	3/22
Germany	0	0	0	0	0	0	3	3

Government variables:

Strength and Technique of Response (2.5) – This variable looks at how the government has typically responded to attacks by various terrorist groups in the past. The harsher and more brutal their response, the higher their score, as this could serve both to enrage and thereby aid the recruitment of future members and also erode popular support of the rest of society. Data was pulled from State Department briefings as well as various press/historical accounts of the past twenty-five years.

Antagonistic or Agonistic Stance (2.5) – Related to the former variable, this score indicates how antagonistic or tolerant a government is of terrorist groups within its borders. The first variable looked solely at the government’s response to a specific attack, while this focuses more on their actions in between the violence. The more provoking and belligerent a government was in these intervening periods, the higher their score for reasons similar to those just given. Data from the 2004 CIA World Factbook was used to determine scores. [This variable’s results were averaged with the former’s in an effort to conserve space and due to correlation amongst the pair.]

Allegations of Torture or Illegal Imprisonment (5) – Countries that torture or illegally imprison perceived enemies without adhering to the edicts of due process are creating a dangerous environment where crimes of retribution by the relations of the victim(s) are encouraged. Countries with existing and frequent reports (weekly or monthly, depending on publication frequency) of torture or illegal imprisonment therefore received 5 points in the model while those with only occasional or relatively minor allegations (once every couple of months) received 3. (Countries with few or no such allegations received 1 point, the lowest risk.) Data for this variable came from the State Department and its human rights reports[6] as well as information from Human Rights Watch[7] and Amnesty International.[8]

Rumored Ties to Group (Funding or Harbor) (5) – Rumored ties of a government to a group in its midst, either through direct funding or safe harbor and lax prosecution (a la pre-Karzai Afghanistan with the Taliban and Al Qaeda), resulted in higher scores (5 points), the thought being that states such as these were fostering more unrest and violence in their society rather than less. Those with no ties and an active policy towards removing members from the country received 1 point while those with moderate ratings of both received 3. Data used to evaluate states here came from the same sources as the Group variables – the Council on Foreign Relations, the State Department, etc.

Regime Instability (5) – This variable, based on data pulled from World Bank reports on governance and stability from 1996-2002,[9] evaluates how likely a government is to fall in the near future, either by “unconstitutional and/or violent means, including domestic violence and terrorism.”[10] The World Bank’s evaluation takes into account data from several surveys and gives percentile scores for how stable a government is. Thus higher levels of stability in their survey translates into lower scores on ours, the rationale being the more unstable a government is, the more prone to violence and attack their country is. (Countries in the 85^th percentile and above received 1 point, those between the 50^th and 85^th received 3 points, and those below the 50^th percentile received 5 points.)

Corruption and Enforcement of Laws (5) – The World Bank has also collected data on governmental corruption, supplementing their assessment with the indicators of “government effectiveness,” “regulatory quality, “rule of law,” and the more direct “control of corruption.” These four indicators paint a broader picture of the type of government in place and have been used in lieu of the more well known Corruption Perception Index by Transparency International. (These surveys use the majority of the sources TI does, and thus are of comparable import.) A higher score here, then, reflects a more inept, unjust, and corrupt government, one more apt to provoke public ire and be unsupported and the target of a possible terrorist attack. (Reflected in sub-50^th percentile scores in at least 2 of the 4 variables – 5 points.) Countries in the 85^th percentile or above in at least 2 of the 4 variables, conversely, were considered the lowest risk and received only 1 point. (Those between the 50^th and 85^th percentile in at least 2 of the 4 received 3 points.)

Table III: Breakdown of Government Score Tabulations

Country	Strength / Technique of Response	Antagonistic / Agonistic Stance	Allegations of Torture / Imprisonment	Rumored Ties to Group (Funding, Harbor)	Regime Instability	Corruption / Enforcement of Laws	Total
Portugal	1	1	3	0	3	1	8
Spain	1	3	3	1	1	1	8
Italy	1	1	1	0	3	3	8
Greece	1	1	1	1	3	3	9
France	1	1	3	0	3	1	8
Germany	1	1	1	0	1	1	4

Infrastructure variables:

Size, Strength, and Security of Borders (5) – The bigger a country is, the more options there are for arms and/or terrorists to cross; the stronger they are, the less likely the chances of a group forming an ideology on and fighting over territory and succeeding; the more secure a country’s borders, the lower the chances of the first scenario being a problem. Thus high scores here reflect any combination of large, weak, and insecure borders (5 points), with higher scores coming for countries exhibiting all three. (Those with moderately protected borders and equal risk to attack received 3 points while those with strong, secure borders earned only 1.) Data was pulled from State Department descriptions and the Federation of American Scientists website.[11]

Size, Strength, and Security of Railways, Airports, and Ports (15) – The more prevalent a state’s air, rail, and sea hubs, the more potential targets a terrorist group has to select from and the more insecure these are, the better their chances of success. Therefore high scores here reflect either a pronounced number or weakness in any of the aforementioned arenas, with higher scores coming for countries exhibiting those characteristics in all of the above. For ports, countries with less than 5 earned 1 point, those with 5-10 received 3, and those with over 10 earned 5 points; for airports, countries with less than 15 medium- or large-sized airports (those with paved runways of longer than 2438 feet) received 1 point, those with 15-50 such facilities earned 3 points, and those with over 50 received 5; for railways, countries with less than 1,000km of track earned 1 point, those with 1,000-10,000km earned 3 points, and those with over 10,000km earned 5. Data was pulled from the CIA World Factbook (2004) to determine scores for size in all three categories[12] while common knowledge, press, and industry reports were used to gauge their security.

Table IV: Breakdown of Infrastructure Score Tabulations

Country	Size/ Strength/ Security of Borders	Size/ Strength/ Security of Ports	Size/ Strength/ Security of Airports	Size/ Strength/ Security of Railways	Total
Portugal	1	3	1	3	8
Spain	3	5	3	5	16
Italy	1	5	3	5	14
Greece	1	5	3	3	12
France	1	5	3	5	14
Germany	1	5	5	5	16

Society (Internal) variables:

Public Opinion of Group or Government (2.5) – To evaluate the strength of probable support for various governments (rather than using popular opinion polls, which we felt were too prone to fluctuation and bias) we used Freedom House’s indicators on “political rights,” “civil liberties,” and World Bank’s indicator for “voice and accountability.” The former two variables are calculated by evaluating how well a country fits a series of questions on a checklist (ten for the first, fifteen for the latter)[13] while the latter utilizes a series of surveys to make its assessment. Our model gives higher scores to those countries with poorer protection of political rights and civil liberties and lower amounts of accountability, the rationale being that these governments will be seen as being less just by the public, enjoy lower levels of popular support and foment more dissent, and thus be more prone to overthrow or attack. (5 points were given for countries with “not free” distinctions in the political and civil rights categories and “voice” scores below 50^th percentile. Those with “partly free” descriptions and scores between the 50^th and 85^th percentiles received 3 points while those that were “free” and in the 85^th percentile or above received only 1.)

Press Opinion of Group or Government (2.5) – This variable relies on Freedom House’s report on press freedom,[14] which looks at how restrictive the legal, economic, and political environments are for a country’s press and rates them accordingly.[15] High scores in this category indicate a press that is more constrained by the government (5 points – “not free” characterization), thus creating a situation similar to the one above – the government is perceived as being more unjust and poorly supported, the restrictiveness fosters aggravation, organization, and dissent, and is thus more likely to experience an attack. Countries whose press was “partly free” earned 3 points and those that were completely free received 1. [This variable’s results were averaged with the former’s in an effort to conserve space and due to correlation amongst the pair.]

Islamic or Religious Fundamentalism (2.5) – Societies that have higher levels of religious fundamentalism, especially radical Islamic ideologies, tend to be more intolerant and foment more violent forms of protest. Thus higher scores in this category reflect higher levels of fundamentalist behavior, in both size and strength, with more points going towards those societies with greater proportions of the population held sway by these groups and to those with Islamic over other forms of fundamentalism. State Department reports were used to make assessments for this variable.

Anti-Americanism (2.5) – This variable looks at whether a group active within a particular country has an ideology based on, or incorporating, anti-American sentiment. Using the same sources as for the other Group variables, high scores here indicate more vociferous levels of anti-Americanism (5 points), the rationale being that groups driven by anti-American thought are usually more aggressive and tend to express themselves violently, earning an equally violent response, and that this type of mindset is particularly magnetic and able to multiply in environments like this once it is breached. (Countries with moderately anti-American groups earn 3 points while those with low/no such levels received 1.) The same sources used to determine all other group scores were used here as well.[16] [This variable’s results were averaged with the former’s in an effort to conserve space and due to correlation amongst the pair.]

Youth Bulge (5) – Countries that have higher proportions of young people, specifically young men, have been shown to be more prone to violence and crime as these individuals are easier to recruit as they are often not employed fulltime, more malleable to various ideologies, and more energetic, and potentially violent, participants in resistance and protest. High scores here indicate a greater proportion of young people age 14 and under to the rest of the population (5 points for over 35 percent), and thus a heightened probability of violent organization or attack. Countries with more balanced amounts (between 15 and 35 percent) earned 3 points while those with extremely low levels of young people (under 15 percent) received only 1.

Table V: Breakdown of Society (Internal) Score Tabulations

Country	Public Opinion of Group / Government	Press Opinion of Group / Government	Islamic / Religious Fundamentalism	Anti-Americanism (Group)	Youth Bulge	Total
Portugal	1	1	1	0	1	3
Spain (ETA/GRAPO)	1	1	3	1	3	6/8
Italy	3	3	1	0	1	5
Greece	3	1	1	5	1	6
France (Alone/With ETA)	1	1	1	0/1	3	5
Germany	1	1	1	0	1	3

Surroundings variables:

Bad neighborhood (5) – Internal societal variables are not the only key influences on the relationship between a state and the resistance or terrorist groups in its midst. Exogenous factors such as the type and tenor of countries nearby must also be considered in assessing a country’s risk. Here, the thesis is that countries that are antagonistic, violent, or host to a particular terrorist group have the potential to spill those contents over to an adjacent country, increasing the threat of attack to those countries. Thus by looking at Infomundo’s Uneasy Neighborhood Index, which takes into account anecdotal evidence and data from years of risk analysis and reporting, we assess how volatile and potentially hazardous the surrounding environment is for a given country, with higher scores going to those places where it is more pronounced. (5 points for those with “poor” surroundings.) Those countries with average neighborhoods received 3 points, while those with safe, calm, and otherwise innocuous surroundings earned only 1.

Member of Relevant International Groups or Treaties (5) – Similarly, countries that formalize those relationships by joining into pertinent groups or treaties (non-proliferation bills, security cooperatives, counterterrorism initiatives, etc.) are thought to be safer because they are both active in containing these key issues and more apt to be helped by fellow members were the need to arise. Thus high scores here reflect a refusal by a country to participate in these activities, indicated by looking at whether they had joined the most basic international conventions – those of the UN.[17] Countries signing less than 5 of the UN’s 12 counterterrorism treaties earned the highest risk rating of 5 points, while those participating in between 5 and 10 earned 3. (Countries signing over 10 of them were considered the safest thanks to their climate of cooperation and earned 1 point.)

Globalization Ranking (5) – Foreign Policy magazine annually ranks countries in their Globalization Index,[18] evaluating them in such areas a technological, political, and economic integration. The thinking here is similar to that of the above -- that countries ranking higher in their index are more connected globally, have more people looking out for them that are more willing and able to aid them, and thus safer than those who are not. Thus high scores here reflect a country that is forced to go it alone, with low levels of interconnectedness and a higher potential for attack. By averaging the scores from 2001-2004, we ranked those placing in the top 15 as the safest in this regard (1 point), those scoring between 15^th and 40^th place as medium risks (3 points), and those scoring above 40^th place as the highest risk areas, earning 5 points.

Table VI: Breakdown of Surroundings Score Tabulations

Country	Bad Neighborhood	Member of Relevant International Groups / Treaties	Globalization Ranking	Total
Portugal	1	0	1	2
Spain	3	1	3	7
Italy	3	3	3	9
Greece	3	3	3	9
France	1	1	1	3
Germany	1	3	3	7

Results / Recommendations

Looking at the breakdown of the results from each of the previous sections and the overall totals listed in Table VII we can see that Spain was deemed to be the country most at risk for a terrorist attack.[19] Their high risk evaluations in the majority of the composites, particularly the Group subvariables, are key weaknesses that make the country susceptible to attack. Spain received two scores in several of the tables, one for each of the established terrorist groups active within its boundaries – ETA and GRAPO – and their tallies were two of the highest in the model. The size and activity of the groups, along with Spain’s large and vulnerable system of state infrastructure, gave the country such high overall risk evaluations and highlighted two key areas for the government to shore up.

Greece also looks to be at a heightened risk for attack, both for its high overall scores and the three terrorist groups operating within its borders.[20] (They scored 50, 44, and 45 points among its three groups.) Its high Government score – the highest of all surveyed – with its instability and corruption ratings, in addition to its equally robust Surroundings tally – tied with Italy for the highest overall – place the country in a state of accentuated risk for attack.[21]

The French case is interesting depending on whether or not you include the terrorist ratings for ETA. The group’s members have frequently sought refuge in Southern France when pursued by the Spanish government and the group has occasionally acted in French territory as part of its claim for the Basque homeland, but has not done so since August 1996, instead focusing the wide majority of its actions within Spain. Therefore it did not make logical sense to give France the full ETA score that Spain received as attack was a distant possibility when considered next to the Spanish case, but it was equally irrational to dismiss it altogether, thus the dual rating. Therefore the probability of attack by ETA in France lies somewhere in between these two numbers and depending on which pole you shy towards either leaves the country as a middling risk with an overall score in the low 30s or a much stronger threat with one in the 50s.

The remainder of those countries surveyed seem to be at a much more diminished risk of attack, all of them scoring in the 30s or below. Portugal, somewhat incredibly, scored a measly 21 points in the model meaning it has almost no risk of being attacked. The country has no terrorist groups active in its borders, including an absence of al Qaeda, it has very little state infrastructure, an almost non-existent Society score, and an equally diminutive Surroundings score. In fact, it held the lowest score in four of the five categories with tallies often half that of the next closest rating.

Therefore the perception of southern European countries being laggards in comparison to the rest of Europe (represented here by Germany and France) and thus at higher risk for attack holds only intermittently here, corroborated by the Spanish and Greek cases, but failing in the Portuguese and Italian ones. Our recommendation to these governments, then, is to target the aforementioned areas of weakness in an effort to provide their citizens with safer societies – Spain should focus on its infrastructure and terrorist groups; Italy and Greece should look at their corruption, press restrictions and freedom of speech; all countries should continue actively targeting terrorist groups, going after their cells and their funding, while shoring up expansive systems of state infrastructure and increasing their international participation. These will be helpful steps in dissipating the terrorist threat and should be implemented immediately.

Several other variables were used in the last stage of analysis to round out our final assessments. These include whether a country supported the war in Iraq,[22] whether a country had a nuclear, biological, and/or chemical weapons program(s)[23] that was active,[24] sizeable,[25] and/or well funded,[26] and whether a country’s political allies were weak and/or scarce.[27] These were used to determine whether the final recommendations should be augmented or diminished and would have been formally included in the scoring were space not an issue.[28]

Table VII: Breakdown of Final Score Tabulations

Country	Group	Government	Infrastructure	Society	Surroundings	Total
Portugal	0	8	8	3	2	21
Spain (ETA/GRAPO)	22/12	8	16	6/8	7	59/51
Italy	3	9	14	5	9	39
Greece (Nov 17/ ELA/RN)	14/8/9	9	12	6	9	50/44/45
France (Alone/With ETA)	3/22	8	14	5	3	33/52
Germany	3	4	16	3	7	33

Limitations / Future Expansions of the Model

Obviously the primary limitation to this model is the level of subjectivity used in assessing the various countries. The dearth of hard, statistical data on things like border security, regime instability, port/airport/railway vulnerability and a host of others places the onus of accuracy squarely on the analysts’ shoulders. Regardless of how high quality these individuals are,[29] the amount of opinion and bias that could creep into the data and unwittingly affect the outcome must be acknowledged.

Even when the numbers came from reputable, well-established sources – Freedom House, the World Bank, the State Department – qualitative assessments were not completely purged from the record. Indeed, more often than not these organizations employed similar methods of analysis to our own, assessing a range of qualitative information sources and affixing a number on the result. And while they arguably have access to bigger, better amounts of data and higher levels of technical expertise than us,[30] the potential fallibility of their numbers must again be pointed out.

Part of the problem lies in the question we’re seeking to answer – who is at risk for a terrorist attack? This is an inherently subjective topic to deal with and if there were hard, easily quantifiable data to use in solving it, predicting the next terrorist attack would pose as much trouble as peeling an orange. Therefore we set out to examine a range of more unorthodox variables – unorthodox and more difficult to quantify, perhaps, but unequivocally influential pieces of the larger problem of terrorism – and to sacrifice statistical precision in the interests of covering a wider expanse of territory. Thus the model was to be an exercise akin to Infomundo’s Wildcard Analysis projects and an attempt was made to evaluate otherwise ignored indicators while instilling order to a situation where only chaos had reigned before.

Future uses of this model – which we still believe to be a viable one in spite of the aforementioned limitations -- should try to incorporate more rigid, quantitative data sets rather than relying so heavily on qualitative assessments of the variables. It should also widen its survey set to a number of countries rather than refine it to six to strengthen the quality of the data and its subsequent analysis.[31] For avoiding analysis on a topic such as this just because it is difficult or impossible to maintain the preferred levels of scientific rigor is a price none of us can afford to pay.[32]

---

[1] Council on Foreign Relations website, Terrorist Groups subsection. November 22, 2004. http://cfrterrorism.org/groups/

[2] Congressional Research Service; Library of Congress. “Foreign Terrorist Organizations.” Published on CRS Web, February 6, 2004.

[3] State Department. “Patterns of Global Terrorism Report -- 2000.” http://www.state.gov/s/ct/rls/pgtrpt/2000/2450.htm; “Patterns of Global Terrorism Report – 2002.” http://www.state.gov/s/ct/rls/pgtrpt/2002/pdf/; “Patterns of Global Terrorism Report – 2002 Europe Overview.” http://www.state.gov/s/ct/rls/pgtrpt/2002/html/19985.htm

[4] Congressional Research Service; Library of Congress.

[5] International Policy Institute for Counter-Terrorism website. November 22, 2004. http://www.ict.org.il . The E-Text Archives were consulted as well. (E-Text Archives website, Politics subsection. November 22, 2004. http://www.etext.org/Politics/Arm.The.Spirit/Guerrilla/Europe/GRAPO/grapo.mini-history.1985.)

[6] State Department. “Country Reports on Human Rights Practices.” February 24, 2004. http://www.state.gov/g/drl/rls/hrrpt/2003/27865.htm

[7] Human Rights Watch website. November 22, 2004. http://www.hrw.org/

[8] Amnesty International website. November 22, 2004. http://www.amnesty.org/

[9] World Bank website, Governance Indicators:1996-2002 subsection. November 22, 2004. http://www.worldbank.org/wbi/governance/govdata2002/.

[10] World Bank. “Government Matters III: Governance Indicators 1996-2002.” Published April 5, 2004. http://www.worldbank.org/wbi/governance/pdf/govmatters3_wber.pdf

[11] http://www.fas.org/main/content.jsp?formAction=325&projectId=22

[12] Central Intelligence Agency. “The World Factbook 2004.” http://www.cia.gov/cia/publications/factbook/index.html

[13] Freedom House website, Freedom in the World 2003: Survey Methodology subsection. November 22, 2004. http://freedomhouse.org/research/freeworld/2003/methodology.htm

[14] Ibid., Freedom of the Press 2004 subsection. http://freedomhouse.org/research/pressurvey/regionaltables2004.pdf

[15] Ibid., Freedom of the Press 2004: Survey Methodology subsection. http://freedomhouse.org/research/pressurvey/methodology2004.pdf

[16] Further expansions of this model should account for anti-Americanism on the part of the state’s leaders or its general public. Since this region, with the exception of France, is largely devoid of such sentiment and the data quantifying its degree would be rather subjective, we decided to focus solely on the anti-Americanism of the groups based on various statements or communiqués they have made in the past.

[17] State Department. “Patterns of Global Terrorism – 2002 Europe Overview.”

[18] Foreign Policy website, Measuring Globalization: Economic Reversals, Forward Momentum subsection. November 22, 2004. http://www.foreignpolicy.com/issue_marapr_2004/countrydetail.php

[19] We didn’t break down the overall scores into degree demarcations as we were only concerned with who was most at risk for attack. (Our limited survey set of six countries and their overall similarity also made this a rather moot point.) Were we to run this model again with a wider survey of countries we would be forced to breakdown the overall scores into different categories. Just judging by the outcome of this incarnation of the model, we would assume that high risk countries would run from around 65-100, medium risk countries from 30-64, and low risk countries from 0-29, but firm separations should not be made until examining the results of a wider survey.

[20] Greece received a similar treatment, earning three scores for its three different terrorist groups – November 17, ELA, and Revolutionary Nuclei.

[21] Security and funding of infrastructure items and the government’s stance against terrorism within its borders were hardened and increased in the runup to this summer’s Olympic games and may diminish their risk evaluations, but data on the work has not yet been released and is therefore not included.

[22] Supported War in Iraq – This variable evaluates whether a country was for the war in Iraq or against it, the thought being that those who supported the war, often in the face of stiff domestic opposition and international condemnation, will be more weakly supported and more prone to be the targets of fundamentalist, anti-American terrorist attacks, thus garnering higher scores here. Sources for this evaluation are surveys of press reports over the past two years and stated positions of various government officials.

[23] Existence of Nuclear, Biological, and/or Chemical Warfare Program – Relying on data from the State Department and NTI (Nuclear Threat Initiative), we determined whether or not a country had created programs in any or all of these categories. The rationale here was that the presence of these programs indicated both how aggressive and potentially violent a country was in the protection of its assets and interests, and also whether a group would have the option, were it so capable and inclined, to steal items from these programs to use for itself. The more of these programs a country had, the higher its score for the effect the just stated reasons could have on fomenting violence within its borders and abroad.

[24] N/B/C Warfare Program Active or Retired – Whether or not these programs are still in existence is an important indicator of both how aggressive and/or belligerent a state may be in response to terrorist attack and how easy it might be for a group to obtain remnants from their demise. (Think Russia, with its massive and incredibly advanced programs from these areas that have since been mothballed and poorly protected, posing a major threat to international security through black market poaching.) Lower scores here reflect the perception that active programs will be better protected than defunct ones. (1 point -- Semi-active programs received 3 points while retired programs earned 5, the highest risk.)

[25] Size of N/B/C Warfare Program – Where the previous variable was merely an indication of whether these programs were in existence, this variable reflects how large a country’s programs are in those areas, the thought being that the bigger and more expansive they are, the more aggressive and potentially violent a state is and the more options a group has to target for theft and personal use.

[26] Funding of N/B/C Warfare Program – Similar to the above, the more well funded these programs are, the better the odds are that they are large and technologically advanced. Higher scores reflect just that breadth and state of advancement.

[27] Weak or Few Political Allies – Countries that are strongly tied to the international community, reflected by a high frequency of communication and cooperation with various nations, are generally thought to be safer from terrorist attack than those who truck it alone as they are more firmly entrenched in the system of mutual protection and observation. Thus high scores in this category indicate a country that is more isolationist or merely more isolated as the result of having fewer, and/or weaker allies in the global community. This makes them more susceptible to attack as there is no one big or bad enough watching their back, so to speak, and they are left to protect themselves by their own devices.

[28] One final variable was used to make our ultimate evaluations: Rumored Ties of Group – If a group was known to have links to other international terrorist groups or governments, be it arms, funding, training, or simple communication, this was an indicator of their global strength and support, thus the higher their score was. Frequency and source were also taken into consideration, with repeated interactions scoring higher and those occurring with similarly active, potent groups scoring more than those with fledgling, impotent groups. (5 points for both scenarios) Those with weak, infrequent, uninfluential ties received 1 point while those with modest, moderate ties received 3.

[29] In this case, particularly exceptional.

[30] Though as Richards Heuer has so notably pointed out in his Psychology of Intelligence Analysis, we must remember that more data doesn’t necessarily mean better data or even better analysis, for that matter. (Heuer, Richards J., Jr. Psychology of Intelligence Analysis. Central Intelligence Agency: 1964.)

[31] Widening the data set allows for assessments that are more applicable to the rest of the world and not made relative to each other. For example, while Italy’s and Greece’s corruption scores may look particularly bad compared to those of France or Germany, when placed next to those of Russia or Romania they look rather trifling indeed. Thus widening the survey can only strengthen the resulting analysis and is highly recommended.

[32] A final limitation that caused the researchers some trepidation was the averaging of certain variable scores in the efforts to conserve space and keep the model’s scoring on a 100-point scale. The decision was made to keep the variables and average them rather than jettison them outright because of their perceived importance and explanatory power. Future incarnations of the model should add variables to expand to a 200-point system or devise an alternate solution that eliminates this need.