terrorists were an existential threat and supported measures that would snuff them out for good. The thing about asymmetric warfare is that one side, by definition, is a lot more powerful than the other. And as the saying goes, the race may not be given to the swift, nor the battle to the strong, but that’s the way to bet.
Though terrorist campaigns have a natural arc that bends toward failure, new campaigns can spring up as quickly as old ones fizzle. The world contains an unlimited number of grievances, and as long as the perception that terrorism works stays ahead of the reality, the terrorist meme may continue to infect the aggrieved.
The historical trajectory of terrorism is elusive. Statistics begin only around 1970, when a few agencies began to collect them, and they differ in their recording criteria and their coverage. It can be hard, even in the best of times, to distinguish terrorist attacks from accidents, homicides, and disgruntled individuals going postal, and in war zones the line between terrorism and insurgency can be fuzzy. The statistics are also heavily politicized: various constituencies may try to make the numbers look big, to sow fear of terrorism, or small, to trumpet their success in fighting terrorism. And while the whole world cares about international terrorism, governments often treat domestic terrorism, which kills six to seven times as many people, as no one else’s business. The most comprehensive public dataset we have is the Global Terrorism Database, an amalgamation of many of the earlier datasets. Though we can’t interpret every zig or zag in the graphs at face value, because some may represent seams and overlaps between databases with different coding criteria, we can try to get a general sense of whether terrorism really has increased in the so-called Age of Terror.199
The safest records are those for terrorist attacks on American soil, if for no other reason than that there are so few of them that each can be scrutinized. Figure 6–9 shows all of them since 1970, plotted on a logarithmic scale because otherwise the line would be a towering spike for 9/11 poking through a barely wrinkled carpet. With the lower altitudes stretched out by the logarithmic scale, we can discern peaks for Oklahoma City in 1995 and Columbine in 1999 (which is a dubious example of “terrorism,” but with a single exception, noted below, I never second-guess the datasets when plotting the graphs). Apart from this trio of spikes, the trend since 1970 is, if anything, more downward than upward.
FIGURE 6–9. Rate of deaths from terrorism, United States, 1970–2007
Source: Global Terrorism Database, START (National Consortium for the Study of Terrorism and Responses to Terrorism, 2010, http://www.start.umd.edu/gtd/), accessed on April 6, 2010. The figure for 1993 was taken from the appendix to National Consortium for the Study of Terrorism and Responses to Terrorism, 2009. Since the log of 0 is undefined, years with no deaths are plotted at the arbitrary value 0.0001.
The trajectory of terrorism in Western Europe (figure 6–10) illustrates the point that most terrorist organizations fail and all of them die. Even the spike from the 2004 Madrid train bombings cannot hide the decline from the glory years of the Red Brigades and the Baader-Meinhof Gang.
What about the world as a whole? Though Bush administration statistics released in 2007 seemed to support their warnings about a global increase in terrorism, the HSRP team noticed that their data include civilian deaths from the wars in Iraq and Afghanistan, which would be classified as civil war casualties if they had taken place anywhere else in the world. The picture is different when the criteria are kept consistent and these deaths are excluded. Figure 6–11 shows the worldwide annual rate of death from terrorism (as usual, per 100,000 population) without these deaths. The death tolls for the world as a whole have to be interpreted with caution, because they come from a hybrid dataset and can float up and down with differences in how many news sources were consulted in each of the contributing datasets. But the shapes of the curves turn out to be the same when they include only the larger terrorist events (those with death tolls of at least twenty-five), which are so newsworthy that they are likely to have been included in all the subdatasets.
FIGURE 6–10. Rate of deaths from terrorism, Western Europe, 1970–2007
Source: Global Terrorism Database, START (National Consortium for the Study of Terrorism and Responses to Terrorism, 2010, http://www.start.umd.edu/gtd/), accessed on April 6, 2010.