The following is part of a series of thought pieces authored by members of the START Consortium. These editorial columns reflect the opinions of the author(s), and not necessarily the opinions of the START Consortium. This series is penned by scholars who have grappled with complicated and often politicized topics, and our hope is that they will foster thoughtful reflection and discussion by professionals and students alike.
Discussion Point: Black Swans and Burstiness—Countering Myths about Terrorism
Terrorism has two characteristics that make it prone to myth making—its black swan nature and its burstiness. Essayist Nassim Taleb defines a black swan incident as one that falls outside the realm of regular expectations, has a high impact, and defies predictions. The term is based on the observation that before they visited Australia, Europeans had assumed that all swans were white; an assumption that at the time was supported (for Europeans at least) by their own experience.
Taleb claims that the coordinated terrorist attacks of 9/11 are a perfect example of a black swan event because they were unexpected, had a huge impact on policy and were difficult to predict. One of the major challenges in responding to terrorism is that a handful of very rare cases can have a disproportionate effect on setting the agenda for the phenomena more generally.
But terrorism also tends to be bursty. Bursty distributions are those that are highly concentrated in time and space. Recent research has shown that diverse phenomena are bursty, including streams of e-mail messages; traffic on crowded freeways; the frequency of forest fires—and the global distribution of terrorism. These two qualities—its black swan character and its burstiness—make responding to terrorism challenging. On the one hand, terrorism is relatively infrequent and hard to predict; on the other hand, when it starts to happen there is a tendency for it to happen in the same place a lot.
The Global Terrorism Database (GTD), maintained by the START Center at the University of Maryland, and now being used for this report by the Institute for Economics and Peace, can help us put these characteristics of terrorism into context. The GTD includes over 104,000 terrorist attacks from everywhere in the world that took place from 1970 until 2011. In this short essay, I will use the GTD to first reflect on some of the myths generated by the incredible impact of 9/11 and then consider some of the policy implications of the bursty nature of terrorism made evident by the GTD.
I will begin by discussing nine myths about terrorism that have been strongly influenced by black swan events like 9/11. I call these myths in the everyday sense that they are conclusions that are fictitious or unscientific. First, 9/11 had an immediate and dramatic impact on levels of public concern about terrorism in the United States and well beyond. Accordingly, many observers assumed that terrorist attacks and fatalities were up sharply in the years just before 9/11.
But in fact, the GTD tells a different, more complicated story. According to our data, terrorist attacks reached their twentieth century zenith not in 2001, but in 1992—just after the collapse of the Soviet Union. Total attacks the year before 9/11 were at about the same level as they had been at in the mid-1970s.
Second, the ubiquity of modern communications systems means that individuals are now continuously bombarded by images of terrorist attacks from around the globe. Consider how many times you have you seen the iconic 9/11 image of a jet plane crashing into the World Trade Center? This blanket media coverage leaves the impression that no location on the planet is safe from terrorism. But in fact, our analysis of the GTD indicates that terrorist attacks are highly concentrated in a relatively few places. For example, the top ten countries in terms of terrorist attacks account for nearly half of all terrorist activities in the world since 1970. Ten percent of the world's countries account for 75% of the world's terrorist attacks.
Third, the devastating impact of 9/11 led many observers, both in the United States and elsewhere, to assume that the United States is the target of an inordinate number of terrorist attacks. However, when we use the GTD to examine the frequency of attacks and the number of fatalities by country, we find that the US ranks about 14th in the world in terms of total attacks and about 16th in terms of total fatalities. And while the US ranks 16th in terms of total fatalities, 90 percent of total US terrorism fatalities since 1970 are accounted for by the four coordinated attacks of 9/11. If these attacks are removed from the estimates, US fatalities from terrorism are similar to fatalities for Canada or Greece.
Fourth, the tremendous impact of 9/11 encourages us to think about terrorism as being mostly about dissatisfied individuals from one country attacking innocent civilians from another country. My colleagues and I recently looked at the attack patterns of 52 foreign terrorist groups that were identified by the US State Department as posing the greatest threat to US security.
Based on GTD data we found that more than 90 percent of the 17,000 attacks carried out by these groups were actually domestic attacks. This means that groups located in, for example, Pakistan, were far more likely to use terrorist violence against non-US targets in Pakistan than they were to attack US targets in Pakistan or to attack the US homeland.
Fifth, because of the seeming irrationality of the al Qa'ida attack of 9/11, it is easy to lose sight of the fact that a large number of terrorist attacks involve fairly rational political disputes over territory. Although there are major differences in terms of their specific orientation, this explains in large part all of the top ten most active terrorist groups of the modern era, including Shining Path, ETA, the FMLN, the IRA, FARC, Hamas, and the LTTE.
Sixth, because of the devastating attack of 9/11, it is easy to suppose that most terrorist attacks are incredibly lethal. However, from the GTD we find that more than half of all terrorist attacks since 1970 involved no fatalities. Many incidents are directed at property: bridges, electric plants, factories. Others attacks are aimed at civilians, but they fail. And in many other cases terrorist groups provide a warning to civilians before striking. This has been a common practice for ETA and the IRA and used to be a common practice for the Weather Underground.
Seventh, the coordinated attacks of 9/11 involved long term planning, split second timing, and an innovative use of existing resources. And the sophistication of 9/11 pales into insignificance compared to the diabolical cleverness of the enemies that Clair Danes, Kiefer Sutherland and other television and media heroes routinely face. These images no doubt encourage us to think that most terrorist strikes depend on sophisticated weaponry. But contrary to the view of terrorism that we commonly get from Hollywood, the vast majority of terrorist attacks rely on unsophisticated, readily accessible weapons.
According to the GTD 80 percent of all attacks rely on explosives and firearms. And for the most part, the explosives used are relatively common, especially dynamite and grenades. Similarly, the most common firearms are also widely available, including especially shotguns and pistols. Fortunately, sophisticated weapons, including chemical, biological, radiological and nuclear weapons, are the rare exception.
Eighth, given the persistence of high profile, long lasting groups like al Qa'ida, the Tamil Tigers or the Irish Republican Army, there is also a common perception that most terrorist groups have long life spans. The GTD identifies more than 2,000 separate terrorist groups. We gauge their longevity by the amount of time from their first strike to their last known strike. We find that nearly 75 percent of the terrorist organizations identified in the GTD last for less than a year. Most terrorist groups are like most business startups—very likely to disappear during their first year of operation.
And finally, the advance planning, originality and destructiveness of 9/11 contributed to the notion that terrorist groups are infallible. We could call this, the myth of the super terrorist. My colleagues and I at the START Center have been involved in several research projects using GTD data which suggests otherwise. For example, in a recent study we used the GTD to examine the targeting strategies of the Armenian Secret Army for the Liberation of Armenia (ASALA)—a very active group based in Turkey.
We were especially interested in ASALA because after mounting a long series of deadly terrorist attacks throughout the 1970s and early 1980s, its attacks abruptly declined. After modeling many possible explanations for this sudden desistance, our conclusion was that the most convincing explanation was a strategic shift by ASALA in its targeting strategy. Before the early 1980s, ASALA was careful to target Turks and avoided non-Turkish and especially Armenian, casualties. But starting in the early 1980s, they became far less discriminate in their targeting methods.
The pivotal historical event in our analysis was an especially brutal attack on Paris's Orly Airport in 1983 that killed eight people and wounding over fifty more. The increasing reliance on random, brutal violence such as the attack on Orly created a polarized and hostile climate among the supporters of ASALA and seriously undermined its legitimacy. In short, ASALA badly miscalculated the impact of its changing strategy on its followers.
So, contrary to our stereotypes based on 9/11 and a few other extraordinary events, most terrorist attacks for the past four decades have relied on readily available, unsophisticated weaponry, and frequently involve few or no fatalities. The typical terrorist group disappears in less than a year and there is ample evidence that terrorists frequently make strategic errors. Attacks were declining just before 9/11 and very few attacks involve disgruntled groups from one country attacking civilians in another country. If 9/11 is a black swan event, why not simply ignore it and go back to business as usual?
A major reason why ignoring terrorism is a risky idea is directly related to its burstiness: when it starts to happen it happens a lot and rapidly. We have many examples of this phenomenon provided by research being conducted at the START Center. Attacks by groups that target the United States follow this pattern so that if we look at attack patterns over the past four decades we see not one single group but three waves of attacks. We see similar patterns when we look at aerial hijackings, suicide bombings, and improvised explosive devices.
There is a rapidly developing literature suggesting that a wide variety of different types of crime and violence may have this bursty quality, including burglaries, robberies and gang violence. My colleagues and I have begun to apply similar logic to the study of terrorism. For example, in a recent study we examined the spatial and temporal distribution of terrorist attacks by the Basque Separatist Group ETA in Spain from 1970 on and the attacks of the Farabundi Marti National Liberation Front (FMLN) in El Salvador from 1980 to 1992.
We were especially interested in measuring what we call microcycles—localized bursts of criminal or violent attacks. We began by classifying 2,000 terrorist attacks attributed to ETA and 3,300 terrorist attacks attributed to the FMLN into a time-space grid. We found that for ETA, 52% of all attacks happened within microcycles that were within two weeks and 5 miles of each other; 60 percent happened within microcycles that were within two weeks and 10 miles of each other. The concentration was even greater for the FMLN: 67% of FMLN attacks happened within microcycles that were within two weeks and 5 miles of each other and 81% happened within microcycles of 2 weeks and 10 miles of each other.
Moreover, we found that compared to other attacks, attacks that were part of microcycles had significantly different characteristics in both countries. For example, compared to other types of attacks, bombings by both ETA and the FMLN were likely to be part of microcycles while assassinations and armed attacks were likely to be isolated events. We also found that whether an attack was part of a microcycle was closely related to its location. For both ETA and the FMLN, attacks on the national capital were more likely to be part of microcycles. While these results are preliminary, they give us reason to hope that analysis of the spatial and temporal patterns of terrorism might help to guide policies on countering terrorism.
Which brings me to a few conclusions. I have argued that policies on terrorism are strongly affected by black swan events and that 9/11 is a good example of such an event: it was unexpected, of great magnitude, and had a huge impact on policy. But in addition, terrorism has a bursty quality. When it is effective in a particular time and place, we get a lot of it rapidly. This last point suggests that it would be foolhardy to ignore the threats posed.
And this is the challenge for contemporary societies raised by terrorism: there are dangers in over reacting but there are also dangers in not reacting. Fortunately, 9/11 has turned out to be a rare event—a black swan. But unfortunately, the threat of sudden bursts of terrorist attacks is likely to be a permanent feature of the twenty-first century.