Abstract
In most contemporary Islamist terrorist plots in the West, we suspect a transnational component. This study attempts to identify and quantify this component. The author has created a data set of Islamist terrorist attacks, and foiled and failed plots in the West as well as against Western targets outside the West, from 1989 to 2008. From similar findings in studies of left-wing terrorism of the 1960s and 1970s, a contagion effect from terrorism outside the West to inside the West is expected. Using ordinary least squares (OLS) regression, this paper finds a very strong relation between terrorism against Western targets abroad and terrorism in the West, with a 2-year time lag. This relation proves robust even when controlling for the effect of the two Gulf wars and the war in Afghanistan. Surprisingly, no statistically significant effect from these three wars can be identified on the level of terrorism in the West. Instead, the wars apparently had a significant effect on the level of terrorism against Western targets outside the West, thereby possibly indirectly affecting the level of Islamist terrorism in the West.
Notes
1. For a further discussion of waves of terrorism, see Kaplan, Citation2007; Koopmans, Citation2004; Rasler & Thompson, Citation2009; Sedgwick, Citation2007.
2. In 2007, Radio Free Europe studied the content of a number of ‘semi-official’ Islamist militant websites documenting the focus on Iraq and Afghanistan (Kimmage, Citation2008). The impact of Desert Storm on Osama bin Laden, the creation of al Qaeda, and Islamist militancy generally is widely documented (e.g., Gerges, Citation2005; Kepel, Citation2002; Wright, Citation2006).
3. Some countries, like the UK, changed their counterterrorism laws in 2000, due to the perception of a change in the level of threat posed by Islamist militants (homeoffice.gov.uk, 2009).
4. Available on author's webpage: www.harrow.ws (databases A, B and C).
5. A number of cases, notably the many poison cases in 2002–03 in Rome, Paris, and London, were made public by police and security services and surrounded by vast media attention, but were largely unsubstantiated (e.g., Cowell, Citation2005). Similarly, a number of cases have been advertised as actual plots but ended in courts with convictions on lesser charges, immigration offenses, recruitment for and financial support to jihadists outside the West, etc. These have also been excluded. A number of cases, particularly in the USA, have been criticized for the (too) active role of civilian agents. However, this study has included all plots resulting in conviction.
6. Algeria, 1992—2002; Somalia from 2002; Iraq from 2003; Afghanistan from 2001; and all cases relating to the Balkan wars.
7. Bivariate regressions have also been carried out with no time lag and lags of 2 and 3 years. All of these are far from being significant at p ≤ .05.