Drone Strikes and Evidence-Based Counterterrorism

The debate over U.S. counterterrorism policy in the wake of the U.S. withdrawal from Afghanistan has taken on greater urgency in light of a U.N. report released earlier this year concluding that “terrorist groups enjoy greater freedom in Afghanistan than at any time in recent history.” 

U.S. counterterrorism defenses are much stronger now, but the United States’ plans also rely on “over-the-horizon” capabilities to counter terrorist threats. These involve developing intelligence sources and operational bases at a greater distance from threat locations than over the past 20 years, in order to monitor groups and potentially conduct kinetic operations against them. Jonathan Schroden recently offered thoughtful suggestions on Lawfare on how the United States might address the challenges in doing this without a fully cooperative partner in the Taliban. He and other commentators assume that drone strikes will continue to play some role in this strategy.

How much is known, however, about what 20 years of drone strikes outside war zones have and have not accomplished? Have they been an effective counterterrorism tool? Have they been a precise “surgical” instrument that hits militant targets while minimizing civilian casualties? Have local populations greeted strikes with gratitude or resentment? Decisions about whether, when and where to rely on strikes should be informed by the United States’ best understanding of the answers to these questions. Yet both supporters and critics often tend to offer generalizations based on anecdotal evidence and implicit assumptions. 

More than 60 empirical studies have the potential to shed light on the impacts of targeted strikes in general and of the U.S. strike campaign in particular. In addition, extensive al-Qaeda correspondence in the repository at West Point illuminates the effects of strikes in the Federally Administered Tribal Areas (FATA) in Pakistan against top al-Qaeda leadership known as al-Qaeda Core (AQC). I review and assess all of this material in my new book, “Drone Strike: Analyzing the Impacts of Targeted Killing.” Because the vast majority of empirical work on the U.S. strike campaign focuses on strikes against al-Qaeda, that is my focus as well.

As the book describes, there are challenges in conducting empirical research that allows for plausible inferences of causality. An increasing number of studies in recent years, however, have used methodological approaches designed to provide more robust bases for causal inferences. Studies vary in their rigor, and their findings are neither uniform nor completely unequivocal. They nonetheless suggest some lessons that can provide a valuable foundation for an evidence-based approach to decisions about the use of targeted strikes. Practical, legal and ethical deliberations about strikes should be informed by these lessons. 

Strike Effectiveness

First, how effective have U.S. targeted strikes been against al-Qaeda? The answer depends on the goal. In the early years when the United States saw itself as engaged in a “war on terror,” the aim may have been to destroy al-Qaeda by eliminating its leadership. That hasn’t happened. The overall group is larger than ever, and empirical research indicates that targeted strikes have had no discernible effect on the number of attacks the group conducts. 

This is not surprising in light of research on other campaigns against militant groups. Targeting leadership is likely to be most successful early in a group’s life, and al-Qaeda was a relatively mature organization when strikes began in 2002. Targeting tends to be unsuccessful against groups that have some degree of bureaucratic organization and operational routine, which makes them less dependent on particular leaders. There also is considerable evidence from al-Qaeda documents that this was the case for the group. 

At the same time, al-Qaeda is structured as a network in which local groups have substantial autonomy to decide when and where to conduct attacks. Research indicates that such groups also are less vulnerable to leadership targeting. The fact that strikes have not affected the activities of al-Qaeda as a whole thus reflects the fact that it has what Jenna Jordan calls a “quasi-bureaucratic” organizational form. 

A more modest U.S. goal is for strikes to lower the risk of a major al-Qaeda attack in the United States. There has been no such attack in the United States since September 2001 nor in the West since the London bombings in July 2005. Is this because of the U.S. strike campaign? The answer almost certainly is no. The main reason is that the United States has substantially strengthened its counterterrorism defenses, intelligence-sharing and law enforcement cooperation activities since 9/11. 

At the same time, however, considerable evidence from AQC correspondence during the time that strikes intensified in the FATA—then the base of al-Qaeda leadership operations—indicates that strikes significantly weakened and disrupted AQC, eventually causing evacuation of what had been a relatively safe haven from which it could plan, coordinate, and train individuals for attacks in the U.S. and the West. While quantitative studies are not precisely structured to identify the consequences of these impacts, qualitative evidence suggests that there is good reason to believe that they contributed to lowering the risk of attack in the United States.

This is because AQC has always been the element of al-Qaeda most deeply committed to attacks against the “far enemy” in the U.S. and the West. It was able to do this from Afghanistan before 9/11 and eventually from the FATA. AQC helped coordinate five major attacks in the West or against Western targets from 2002 to 2005, with a sixth attack planned for London two weeks after the July 7, 2005, attack there. The July 21 attack was unsuccessful only because of improper detonation. In addition, AQC organized six attacks that were thwarted from 2006 to 2013, including one that would have detonated liquid explosives in six passenger aircraft over the Atlantic. 

These projects were consistent with AQC’s insistence that the first step in establishing Sharia law in Islamic countries is, in the words of former al-Qaeda leader Osama bin Laden, “to continue our direct attrition against the American enemy until it is broken and is too weak to interfere in the matters of the Islamic world.” Similarly, bin Laden wrote to a lieutenant, “Given that the difference of the impact of attacks against the foes inside or outside of America is substantial, we need to confirm to the brothers that every effort that could be spent on attacks in America would not be spent outside of it.” Bin Laden’s successor, Ayman al-Zawahiri, has continued to emphasize the importance of focusing on the far enemy. 

Research indicates that the intensification of strikes in the FATA beginning in 2008 eventually thinned the ranks of experienced and capable leaders, and prompted curtailment of communication and movement that impaired their ability to coordinate operations with members of the larger network. Bryce Loidolt’s extensive analysis of al-Qaeda correspondence indicates that the first element of al-Qaeda to feel these effects was the unit responsible for conducting external attacks outside the Afghanistan-Pakistan region. Al-Qaeda began moving members from the FATA in 2010, and bin Laden authorized the complete evacuation of the area in January 2011.

The result was the weakening and dispersal of the element within al-Qaeda that was most focused on attacks in the United States and the West. The network has grown since then through affiliation with a variety of local groups whose primary concerns are their own local campaigns. The percentage of attacks attributable to AQC has declined precipitously over the past 15 years or so, while attacks conducted by other parts of the network have increased. AQC has not organized a successful attack in the West since 2005, nor even any plots that have been thwarted since 2013. It is therefore reasonable to conclude that strikes against AQC in the FATA contributed to reducing the risk of attacks in the United States, even if other measures may have been more significant.

Another plausible goal is to reduce terrorist attacks against U.S. persons and allies in countries in which al-Qaeda is most active. Strikes that do this also may weaken al-Qaeda affiliates’ ability to conduct attacks in the United States and West. Studies generally, although not uniformly, indicate that strikes can reduce the number of terrorist attacks and fatalities from them for one to four weeks in locations where strikes occur. Research on strikes in Pakistan suggests that the ability of strikes to achieve these local effects requires an ongoing intensive campaign of surveillance, intelligence gathering and persistent strikes. Asfandyar Mir and Dylan Moore emphasize the crucial importance of “local partner capacity and cooperation” in this process.

Effects may continue as long as strikes do, but strikes by themselves will not necessarily permanently weaken a terrorist group. They may, however, reduce violence enough to pursue initiatives that could do so. Several observers maintain that these initiatives include local governance and security reforms, and the creation of meaningful economic opportunities, in order to lessen the appeal of extremism. While the potential of such initiatives seems plausible, there are considerable challenges in pursuing them and little rigorous research that tests their efficacy. 

Research conclusions on terrorist attacks in retaliation for strikes are mixed. Some studies find attacks associated with strikes, while others do not. Some evidence suggests that strikes affect the timing but not overall number of attacks, as terrorists may seek to conduct an attack sooner than planned after a strike in order to demonstrate resolve to the local population. Interestingly, a substantial body of research finds no relationship between civilian casualties and attacks, belying the claim that attacks are based on a desire to avenge such casualties.

Impacts on Civilians

A full assessment of drone strikes requires analyzing their impacts not only on terrorist groups but also on civilians. First, the best data available from nongovernmental organizations refute the claim that drone strikes consistently kill more civilians than militants. They also indicate declining rates of civilian casualties in the past several years. New America estimates that from 2002 to 2012, some 11.2 percent of casualties were civilians, while the Bureau of Investigative Journalism’s estimate is 23.3 percent. From 2013 to 2020, New America’s estimate is 3.5 percent and the Bureau of Investigative Journalism’s is 4.4 percent. 

In May 2013, President Obama announced guidance that required near certainty of no civilian casualties for strikes conducted outside of areas of active hostilities. Shyam Raman, Paul Lushenko and Sarah Kreps suggest that the guidance had been in place for some time before it was announced, and they find a decline beginning in July 2011, but I use 2013 as a helpful point of demarcation. 

At the same time, the United States for a considerable period denied causing any civilian casualties and has consistently underestimated them compared to credible estimates from other sources. The failure of the United States to meet its own standard of near certainty of no civilian casualties reflects the fact that drone strike precision is not simply a function of technology but also of the organizational processes that are put in place to deploy it. As experts such as Larry Lewis have noted, minimizing civilian casualties requires collecting accurate data on casualties, aggregating this data in order to identify root causes of casualties, disseminating lessons across all relevant U.S. entities and revising operational guidance in ways that incorporate these lessons. The United States will fail to realize the full potential of drone strike precision unless it institutionalizes these measures. 

Second, while there has not been rigorous empirical research on the subject, there is reason to believe that strikes at times have caused property damage that can have devastating consequences in impoverished areas. Among these consequences may be loss of livelihood and loss of shelter for several families who may live together in a single dwelling. Some investigations also report psychological distress from the persistent presence of drones, but this is hard to disentangle from the trauma resulting from living in areas beset by conflict and terrorist groups’ own infliction of violence. While there are formidable challenges to conducting sound empirical research on these impacts, there is an urgent need for it.

Next, research consistently finds that there is strong local opposition to drone strikes in the areas in which they occur. At the same time, the most rigorous research does not support the claim that this translates into greater support for terrorist groups or increases in terrorist recruitment. Residents also harbor considerable resentment toward terrorist groups in their areas. Opposition to strikes nonetheless can impair the effectiveness of U.S. counterterrorism and counterinsurgency operations and needs to be taken into account in a full analysis of the impact of targeted strikes. 

Looking Ahead

What do these findings suggest about the potential future role of targeted strikes outside war zones as a U.S. counterterrorism tool? 

First—with one important caveat—continued targeting of AQC may not appreciably reduce the risk of attacks in the United States. Targeting success in the FATA reflected the distinctive importance and influence of AQC at that time, and the resources it had to plan, organize and attempt such attacks without needing to rely on affiliates to do so. AQC has been more dispersed since its ejection from the FATA and now is able to provide general guidance but few other resources to induce affiliates to attack the far enemy. 

The caveat is that this assessment could change if AQC regains a safe haven in Afghanistan from which it could coordinate and train persons for attacks in the U.S. and the West. In this case, it would be able to resume directly organizing attacks against the far enemy rather than attempting to exhort affiliates to do so. As Schroden and others emphasize, however, the United States would not be able to draw on the kind of intelligence and law enforcement assistance that it was able to in Pakistan, nor on bases for drone operations as near to their targets. These deficits will make it more difficult to locate targets and will increase the risk of civilian casualties. This suggests that even though the reestablishment of an AQC safe haven in Afghanistan would be a serious concern for the United States, it may not be feasible for strikes to play as prominent a role against the group as was the case in Pakistan. If used, they will need to be complemented by extensive other counterterrorism measures. 

Second, as al-Qaeda expands and further decentralizes, experience indicates that local affiliates may acquire their own ability to plan and conduct attacks in the United States. While al-Qaeda in the Arabian Peninsula was once regarded as a serious threat to the United States, for instance, some believe that Al-Shabaab now poses the greatest threat. The research described above indicates that a sustained strike campaign against a group can reduce its ability to conduct local attacks, which may weaken its capacity to conduct attacks outside its region.

There are two limitations of using targeted strikes as a counterterrorism tool outside of a war zone, however. The first is that achieving this result likely will require intensive intelligence gathering and surveillance, and persistent strikes for an extended period. This could cause local population resentment of the United States that undermines its efforts to establish stable regimes in these countries that might be able to undertake reforms to counter the underlying appeal of extremism. In addition, if the United States does not institutionalize efforts to minimize civilian casualties, an intensive strike campaign is likely to cause them, which is contrary to the United States’ own standard of near certainty of no such casualties in areas outside of active hostilities. 

The second limitation is that, without fundamental governance reforms, the most that strikes may be able to accomplish on the local level is what Israel calls “mowing the grass.” That is, an intensive strike campaign could weaken a group as it becomes a threat to the United States—with recognition that this effect may be temporary and it may well be necessary in the future to conduct further strikes. More controversially, strikes could be used to dissuade a group from aspiring to attack the United States. As the departure of the al-Nusra Front in Syria from al-Qaeda indicates, for instance, the prospect of being the target of U.S. strikes may lead some groups to eschew any effort to attack the United States and focus instead on local conflicts. 

Strikes thus could help ensure that local groups have only local ambitions. In this way, strikes might be part of an ongoing low-level use of force, aimed at avoiding the need for the kind of intensive military campaigns that were central to U.S. counterterrorism efforts for 20 years. It would reflect what Daniel Byman calls the “good enough doctrine,” which involves learning to live with, rather than attempting to eradicate, terrorism.

This doctrine has much to commend it. Using strikes to pursue it, however, raises legal issues, depending on how the United States articulates its policy. If the Biden administration reinstates the original requirement that strikes outside areas of active hostilities be against “continuing imminent threats to US persons,” there is likely to be continuing debate over the U.S. conception of what constitutes an imminent threat that justifies a strike. If the goal is to prevent a group from ever becoming strong enough to pose an imminent threat, the United States will be subject to the charge that this use of force for a preventive, rather than preemptive, purpose is inconsistent with international law. 

The evidence indicates some successes, some failures and some mixed results from the U.S. drone strike campaign over the past 20 years. Some lessons are clearer than others. The crucial point, however, is that decision-makers need to engage with this material as they deliberate on the role of strikes in order to develop a genuine evidence-based counterterrorism policy.