Editor’s Note: This is a guest post by Graig R. Klein. Graig is a doctoral candidate in the Department of Political Science at Binghamton University. His research focuses on domestic conflict, protest, and terrorism. This post is based on his article entitled “Ideology Isn’t Everything: Transnational Terrorism, Recruitment Incentives & Attack Casualties,” which is forthcoming in Terrorism and Political Violence.
Since the al-Qaeda attacks on September 11th, 2001 and the subsequent War on Terror, much of the media, policy makers’, and, academics’ attention has focused on the increase in religious motivated terror groups and attacks since the 1990s. Prior to 1993, there were approximately 45 active religious terror groups; by 1994, there were at least 111 active religious based terror groups. This trend overlaps with an increase in casualty rates since the 1990s. Correlation is clear, but causation continues to be debated. Religious based terror groups may exploit cultural identity differences, be less risk-averse, accepting (or seeking) death and martyrdom, public desensitization to attacks or the increased ad-hoc nature of groups.
These arguments and many more are empirically supported, but miss an important piece of the puzzle. In an article forthcoming in Terrorism and Political Violence, I contend that political mobilization goals are an important determinant of terrorists’ casualty rates per attack. Variation in the target, in terms of transnational or domestic, strongly influences the level of lethality or destruction of an attack. Since attacks, in part, aim to attract and mobilize new supporters, large-scale attacks that risk high casualties within a population a group seeks to mobilize are counterproductive. Social mobilization and civil conflict scholarship clearly demonstrate that indiscriminate violence encourages apathetic civilians to turn against the perpetrators of violence. There is little reason to think the same dynamics unfold between civilians and terror groups. Terror groups need to avoid the negative consequences of indiscriminate violence if they are to fruitfully mobilize supporters. This leads domestic attacks, when compared to transnational attacks, to be less destructive in terms of casualties per attack.
On the flip side, to think about this relationship in context of transnational attacks, a higher casualty rate per attack becomes beneficial. Transnational attacks typically occur outside a group’s potential recruitment base and thus negative recruitment incentives do not mitigate the terror group’s desired level of violence. In fact, if terror groups are successful in provoking a transnational target into an aggressive counterterrorism response, the benefits of an attack increase. The counterterrorist becomes the source of indiscriminate or illegitimate violence, which in turn, reduces a terror group’s recruitment costs as the violence pushes civilians towards the outstretched arms of the terror group. Terrorist leadership exploits desires for vengeance as the group’s activities provide an opportunity to satiate such a desire.
The differences in the costs and benefits of acting within one’s potential recruitment base and outside of it influence a terror group’s planned level of violence. Transnational attacks do not target one’s potential recruits and supports, and, from the terrorists’ perspective ideally provoke indiscriminate counterterrorism, thus the level of violence can be maximized with minimal punishment.
Since terror groups attack both national and transnational targets, a binary distinction of groups is insufficient. To empirically test the implications of my argument I distinguish individual attacks as domestic and transnational. Domestic terrorism is defined by commonality between victims, perpetrators, audience(s) and geographic (country) location of an attack; whereas transnational terror attacks are less discriminant and entangle victims, institutions, audiences, policies or perpetrators of different countries.
I test my argument on the universe of terrorist attacks from 1998-2005. The data come from Piazza’s 2009 article and include 4,694 attacks. Using the recorded information on terrorists’ nationality, nation targeted and the venue country, I create a binary measurement indicating whether the attack was transnational or domestic. Attacks are considered transnational if any of the three are different from the others; 1,459 events are coded as transnational (31.1%). I include measures of terror group ideology and goal structure, al-Qaeda affiliation, freedom of press, group competition, population density and control for the September 11th and Kenya/Tanzania 1998 U.S. embassy attacks as these are tremendous outliers in terms of casualties per attack.
As expected, I found that transnational terror attacks do produce more casualties per attack than domestic attacks. On average, transnational attacks are predicted to result in approximately 9 more casualties than domestic attacks. We must keep in mind that nearly 49% of the terror events in my sample resulted in 0 casualties and 86.4% had 9 or fewer casualties. My findings are illustrated in the figure.
Holding all else constant, we see that casualties per attack are significantly different, at 95% confidence, between domestic and transnational groups when there are less than 35 competing groups (89% of the sample have 32 or less, 11% have 61 or greater). Additionally, the declining trend in casualties across the number of competing groups adds additional support to my argument that political mobilization goals are an important behavioral incentive motivating variation in terror attacks; as competition for supporters increases between groups, the need to prevent counterproductive violence increases.
When acting transnationally, terrorists do not need to restrain the size or subsequent victimization of an attack. While domestic terrorism does result in a higher number of total casualties across time, transnational attacks result in a higher casualty rate per attack. Direct implications may be limited to 1998-2005, but the results do point toward the need to consider increased terrorism casualty rates beyond religious orientation or Islamist fundamentalism.