The Management of Savagery: Policy Options for Confronting Substate Threats

This chapter is intended to do two things. First, it categorizes substate actors and the threats they pose to the United States and its interests. That context provides important perspective. Interests in that context are broader than direct threats to the United States and include threats to allies along with an interest in limiting conflict in critical regions, such as the Persian Gulf. That definition of interest is roughly equivalent to the definition of U.S. interests that Barry Posen and Andrew Ross have presented as a grand strategy of “selective engagement.”²

Second, the chapter provides a brief overview of the policy tools the United States has at its disposal to combat those threats. It begins by focusing on the development of tools for the discrete and discriminate use of force, such as drones and special operations forces. Although those tools are important and provide options other than major ground or air campaigns, overreliance on them is expensive and often counterproductive. The United States must give additional emphasis to other tools, in particular tools for covert action and unconventional warfare and for financial intelligence and sanctions. The United States must be more willing to rely on allies to handle some of those threats, and it must also have a better integrated long‐​term approach to developing the capacity of those partners. Finally, the United States must recognize that the threat posed by substate actors is one that must be managed rather than eradicated.

The three main categories of substate threat can conveniently (if not entirely accurately) be labeled “rebels,” “terrorists,” and “bandits.” Each has distinct goals that shape the threat they pose to the United States and its interests. In addition, each has a relationship to both states and to ungoverned territory within states that conditions the threat they pose. “Rebels” are focused on changing the political order in states other than the United States through the use of force or other extralegal means.³ Whether taking the form of mass uprisings or military coups, rebels primarily pose an indirect but often substantial threat to the United States and its interests, principally in the form of threats to allies. Those allies are often part of the status quo that rebels seek to change. Alternately, rebels may be seeking to prevent the establishment of a new status quo by the United States. An example of the former is the overthrow of the shah of Iran in 1979, while an example of the latter is the Taliban in Afghanistan after the U.S. invasion in 2001. “Terrorists,” al Qaeda being the most notorious, seek to change the political behavior of the United States by using force directly against the United States and its interests.⁴ Changing political behavior is both more and less difficult than the goals of rebels. It is easier in that the goal of changing behavior (i.e., a particular U.S. policy) is easier than changing the order that generates those behaviors (i.e., the U.S. government). Yet it is more difficult as terrorists must rely on the coercive use of violence to punish, as well as threaten to punish further, in order to change behavior. Coercion is always a tricky business, even for states capable of inflicting enormous punishment, as the mixed record of U.S. efforts to conduct coercive bombing campaigns demonstrates.⁵ Rebels, though having more ambitious goals, are typically able to employ more direct use of brute force to change the status quo.

Terrorists and rebels also have different geographic constraints. To paraphrase al Qaeda’s own characterization of the issue, terrorists are focused on fighting “the far enemy,” the remote United States. In contrast, rebels are focused on “the near enemy” (the state against which they are rebelling).

In contrast to terrorists and rebels, “bandits” (including pirates and traffickers in illegal goods) have primarily an economic rather than a political motive for the actions they take. Although bandits may lack a major interest in politics for its own sake, they do have a stake in politics as it affects their operations. However, the distinction as to motive is important in considering how policy options can be developed, as discussed in more detail later.

Bandits can pose both direct and indirect threats to U.S. interests. Direct threats are primarily the use of force against U.S. interests either to extract resources (such as through robbery and piracy) or to defend illegal operations (such as smuggling). Indirect threats emerge from externalities associated with those actions, whether in the form of epidemic drug abuse in the United States or the undermining of the political order in foreign countries through endemic corruption and collusion with bandits. The current conflict between drug cartels and the Mexican state is an example of both direct and indirect threats, as the cartels are directly involved in narcotrafficking and the resulting violence in the United States, but they have also undermined the Mexican state through bribery and infiltration.⁶ The cumulative threat posed by those three groups of substate actors has four salient features. First and foremost, the threat is not existential for the United States absent the emergence of effective domestic rebellion. Even the unlikely nightmare scenario of a terrorist organization’s acquiring a nuclear weapon and detonating it in a major U.S. city would not destroy the United States as a country. That is an important consideration in determining policy responses to substate actors.

Second, not all substate violent actors are in fact threats. Banditry of one form or another is extraordinarily common across the globe, but the vast majority poses no direct threat to the United States and its interests, and it poses minimal indirect threat. Likewise, rebellion is fairly common in many parts of the world but may pose only a minimal threat to the United States.

All things being equal, for example, the Taliban in Afghanistan do not pose a major threat to U.S. interests because they are principally concerned with controlling Afghanistan. The United States did not have significant interest in Afghanistan between the end of the communist regime in 1992 and 2001. It was the relationship between the Taliban and al Qaeda, terrorists who did pose a threat to the United States, which changed the U.S. calculus after September 11, 2001.⁷

Policymakers should realize that the threat posed by the Taliban is still not intrinsic to the group but rather is based on its relationships to other groups who might pose a threat to the United States. For example, the Obama administration’s two main arguments for U.S. involvement in Afghanistan to combat the Taliban were principally about al Qaeda and also the potential for other groups (such as the so‐​called Pakistani Taliban) to use Afghanistan as a base to threaten Pakistani stability (and nuclear weapons). It is those relationships between the Taliban and other groups that actually threaten U.S. interests. If those relationships can be shown not to pose a major threat, then, by extension, the Taliban is not a significant threat.⁸

That distinction may seem obvious, but it is often lost in the discussion of the threat posed by substate actors. It can also be difficult to draw a firm line when substate actors threaten U.S. allies. If militants in Yemen principally threaten Saudi Arabia rather than the United States, is that significant enough that the United States should be concerned?

A third feature of the threat posed by substate actors to keep in mind is that although the three categories are conceptually distinct, the boundaries between them are often very blurred in reality. Some groups may fall into more than one category. Al Qaeda, for example, has goals that involve not only changing U.S. behavior through coercion (terrorism) but also changing political orders in the Muslim world through brute force (rebellion). The FARC (Fuerzas Armadas Revolucionarias de Colombia) in Colombia are (or at least were) primarily rebels, but they rely heavily on banditry (in the form of narcotrafficking and kidnapping) to pay for their rebellion. Groups may also establish cooperative relationships, as exemplified by the Taliban and al Qaeda. Yet conceptual clarity is important for designing the appropriate policy response. Policy options that might be useful for combating terrorists, such as the discrete and discriminate use of force discussed subsequently, may have limited utility in dealing with bandits. Indeed, misapplication of policy options may generate new threats, because bandits might in some circumstances become rebels. I will return to that point in the next section, which is about policy options.

Policymakers must, therefore, carefully focus on the essence of substate threats. For example, despite al Qaeda in Iraq’s adoption of terrorist methods and reliance on criminal activity for finance, its fundamental goal was (and is) to change the political order in Iraq. That goal has recently expanded to include Syria, but it remains primarily about changing the Iraqi political order rather than coercing more distant foes or making money.⁹

Finally, none of the three categories is a novel threat to the United States and its interests. Any policymaker of the late 19th century would have been equally familiar with those types of substate actors. In 1886, for example, the United States was fighting the Apache in the West (rebels against the new U.S. status quo) when a terrorist attack in Chicago’s Haymarket Square captured national attention. Only a few years earlier, the public was captivated by the exploits of various bandits, most notably Jesse James and his compatriots (a group that had previously been rebels fighting against the U.S. government during the Civil War).¹⁰

That historical perspective is important because it indicates that the threat posed by those groups, though changing in form and extent, is likely to endure for the foreseeable future. Policymakers must be appropriately modest in attempting to combat those threats because they are unlikely to be eliminated. Instead, they must be managed. Managing them requires an ongoing effort that is carefully calibrated with regard to costs and benefits rather than a “war” to be won. The psychological effect of the 9/11 attacks blinded many to that reality, but after costly wars in Iraq and Afghanistan, there is an encouraging shift back toward that perspective (though more from exhaustion than wisdom). The historical perspective also underscores that although the threat posed by substate actors is real, it is far from existential.

However, there are a few ways in which the diffusion of technology and techniques has made the substate threat more potent than in previous decades, if still far from existential. Three of those technologies and techniques that have diffused over the previous two decades are modern infantry weapons, improvised explosive devices, and suicide bombing. Without overemphasizing the increase in the threat, it is still important to note.

Modern infantry weapons may not seem novel. Substate actors have always had some access to those weapons. What has changed, however, is that it is now much easier for rebels, bandits, or terrorists to acquire those weapons without state support from sources as diverse as looted state arsenals to international arms dealers.

The access of substate actors to modern infantry weapons has increased for two principal reasons. The end of the Cold War was a major driver, opening up arsenals of the Warsaw Pact to the highest bidder.¹¹ The resulting dislocation in both arms manufacturers and members of the security services created strong incentives to sell both weapons and expertise.¹²

In addition, new intermediaries appeared to provide the link between postcommunist sellers and new buyers. The paramount example of the post‐​Soviet arms dealer is Viktor Bout, a former Soviet military officer who was so successful that the movie Lord of War was at least partly based on his operations. Although Bout has been arrested, similar networks no doubt continue to flourish.¹³

In addition to the end of the Warsaw Pact, the past two decades have seen the collapse of several large, well‐​armed states. Beginning with Yugoslavia and continuing through Iraq and Libya, the end of those states has been a bonanza for substate actors. In Yugoslavia and Iraq, the outgoing weapons fueled civil wars, while with Libya, those weapons have spread across the region to arm groups from Mali to the Gaza Strip.

The second new capability available to substate actors is the improvised explosive device (IED). Though explosive booby traps and mines are not new, the easy availability in the 21st century of solid‐​state electronics combined with the proliferation of both military and homemade explosives has greatly increased both the number and lethality of those devices.¹⁴

The United States has sought to combat IEDs with a host of measures, from electronic jammers to heavily armored vehicles, resulting in an arms race of IEDs vs. countermeasures. The countermeasures cost millions of dollars per unit in many cases, whereas the IEDs typically cost hundreds of dollars, or at most thousands. The result is the ability for substate actors to impose disproportionate costs on states through the use of IEDs.

Finally, the use of suicide bombing has spread over the past two decades. From Africa to South Asia, the suicide bomber has become a signature weapon of terrorists and rebels. That use is perhaps most striking in Afghanistan where suicide bombing was not a rebel tactic despite the brutality of the rebels’ decade‐​long war against the Soviets and their Afghan allies.¹⁵

The causes of that diffusion are contested, as suicide bombing obviously involves a more visceral commitment to a cause than using modern infantry weapons or emplacing IEDs. One argument is that occupation by those perceived to be foreigners is the major driver of that diffusion.¹⁶ Others highlight that suicide bombing has diffused simply because it is highly effective, and terrorist organizations, having seen its efficacy demonstrated, have altered their recruitment and training to emphasize the creation of personnel willing and able to carry out such attacks.¹⁷

Yet whatever the underlying reason for the diffusion, it has made a significant difference in the capability of substate actors. Most notable is the effectiveness of suicide bombers against heavily defended targets. For example, in December 2012, a suicide attacker critically wounded Asadullah Khalid, the head of Afghanistan’s intelligence agency despite his elaborate security precautions.¹⁸ The willingness to commit suicide as part of an attack was also critical to the success of the September 11, 2001, attacks on the Pentagon and World Trade Center, as well as the attack on the USS Cole the previous year.

The result of those changes in substate actors’ capabilities is a substantial increase in their ability to impose costs on states, even those as powerful as the United States. As Barry Posen explains, those technologies allow substate actors to maintain “contested zones” where state power is effectively (if not always successfully) challenged.¹⁹ From the Battle of Mogadishu in 1993 to the downing of a U.S. helicopter carrying special operations forces in Afghanistan in 2011, substate actors have demonstrated their ability to impose costs on even elite U.S. units in a variety of contexts.²⁰

However, it should be equally clear that none of those changes in their capabilities have fundamentally enabled substate actors to pose anything like an existential threat to the United States. Indeed, despite the diffusion of those techniques and technologies, it has been difficult for terrorists to use them in attacks on U.S. soil. Neither the Boston Marathon bombing in April 2013 nor the failed Times Square bombing in May 2010 made use of the more sophisticated IED designs common in Iraq and elsewhere.²¹

The increased threat from substate actors is, therefore, principallyto U.S. interests abroad and particularly to the U.S. ability to project power. That threat is most acute in the case of large‐​scale U.S. military intervention. In Iraq and Afghanistan, the U.S. military has suffered significant costs from rebels and terrorists, which is striking considering both the increase in U.S. power following the end of the Cold War and the lack of a major state sponsor for those actors equivalent to the communist bloc during the Cold War. Although Pakistan has been unhelpful in curtailing the Afghan Taliban, it has not provided anything like the level of support that North Vietnam and its communist allies provided to rebels in South Vietnam during the 1960s.

Yet in parallel to the increase in substate actor capabilities, there have been significant changes in U.S. capabilities. A major shift has been the emergence of options for the discrete and discriminate use of force. “Discrete” means the effective use of force apart from a large‐​scale military campaign. “Discriminate” refers to the use of force against a very limited target set, indeed in many cases against single individuals with limited collateral effects. That option has made the use of force more effective for combating substate actors but also has made it more attractive, which is unsurprising given that much of the capability for discrete and discriminate use of force was developed precisely to make the use of force more attractive. This section briefly describes the emergence of that capability.

After Vietnam, both U.S. elites and the U.S. public were extraordinarily reluctant to use force for fear of another quagmire war. An example of that reluctance was the U.S. response to the Iranian hostage crisis in 1979. Both the elites in the Carter administration and the public broadly opposed the use of force. For example, as cited by Jack Snyder and Erica Borghard, one poll in November 1979 indicated “70% of respondents opposed a U.S. policy of bombing the main Iranian oil terminal if Iran put the U.S. hostages on trial, and 65% disapproved of a rescue attempt.” As time passed, the public became somewhat more hawkish on the use of force, but that support remained limited.²²

The United States after Vietnam thus frequently faced a stark choice: either commit to a major conflict—for which almost no one had the stomach—or simply take no action. The Reagan administration spent the 1980s grappling with that reality. Sometimes, the administration was able to use force against a highly peripheral target (the invasion of Grenada or air or naval action against Libyan and Iranian targets). Yet at other times, it was forced to retreat rather than confront a challenge (the withdrawal from Beirut after the U.S. Marine barracks bombing).²³ However, Vietnam and subsequent events drove changes in the way the United States generates and uses the military that allow more limited use of force. There are three critical changes: (a) the movement from conscription to volunteer military forces, (b) the emergence of precision bombing and supporting capabilities, and (c) the development of vast and robust special operations capabilities. Together, they significantly changed how the United States can and does use force.

The first of those changes, the shift to a volunteer military, was a direct outgrowth of the Vietnam War. American attitudes toward conscription have historically been ambivalent, at best, but the conflict in Vietnam brought hostility toward the draft to a crescendo. That hostility was most noticeable among those who were of draft age, and, as recently demonstrated by Robert Erikson and Laura Stoker, a high likelihood of call‐​up for service in Vietnam had profound and enduring effects on the political beliefs of young men, even if they did not ultimately serve.²⁴

Following several years of study, the Nixon administration formally ended the draft in 1973. It was replaced by a system that relied on attracting volunteers through a combination of incentives, including training and education, and opportunity (e.g., to “see the world”). The new system had initial problems, but within a decade had begun to produce a stable, high‐​quality military force.²⁵

Most notably, the fact that all service members were professional volunteers allowed the U.S. military, particularly the army, to take full advantage of a variety of new technologies and training techniques to enhance military performance. Those technologies included not only the emerging precision‐​guided munitions, discussed next, but also the integration of computer technology into a variety of traditionally “low‐​tech” systems, such as tanks and armored personnel carriers. The training techniques were exemplified by the National Training Center at Ft. Irwin, California, which used sensors across a wide area to enable highly realistic training.²⁶

Even more important than the all‐​volunteer military’s effects on military effectiveness is its effect on the domestic political constraints on the use of force. As noted, resistance to the draft peaked during the Vietnam War in large part because those at most risk of being drafted did not want to go to war. Resistance to the draft and resistance to the use of force thus became intertwined.²⁷

However, with the all‐​volunteer force, there was no need for young men (and women) who did not want to go to war to protest the war itself. They could simply ignore a war, and it, in turn, would ignore them. The same was true of their family members, friends, and political representatives. Those who volunteered for service, in contrast, did so accepting the possibility that they would be deployed to war. Thus, with a few exceptions, there was little protest from military families about the war in Iraq or Afghanistan.²⁸

The result of that shift has been remarkably minimal public protest of all sorts of the wars in Iraq and Afghanistan. Public protest of the Iraq War peaked in September 2005 with a rally in Washington that drew more than 100,000 protesters. Subsequent rallies have been smaller, with a March 2010 rally in Washington drawing 10,000 or fewer.²⁹ Moreover, the wars in Iraq and Afghanistan are the most massive, enduring U.S. uses of force, particularly ground force, since Vietnam. The more discrete and limited uses of force, from bombing Libyan forces to drone attacks in Pakistan and counterterrorism operations in Yemen, have produced even less public protest.

Those latter operations have been made possible by the second major change in U.S. military power: the emergence of precision bombing. More precisely, those operations are enabled by precision‐​guided munitions, which enable incredibly accurate bombing from a relatively high altitude, combined with new systems for intelligence, surveillance, and reconnaissance that allow timely and accurate targeting. As a result of those advances, airpower, although far from omnipotent, is vastly more capable than even two decades ago.

Perhaps unsurprisingly, those technologies were born during the Vietnam War. In the Rolling Thunder series of air operations intended to coerce North Vietnam to end its support for insurgency in South Vietnam, the United States attempted to destroy many relatively small high‐​value facilities (power plants, bridges, etc.). Those targets proved difficult to destroy with existing bombing techniques, requiring multiple attempts in the face of often‐​intense air defense and concomitant loss of aircraft.

As a result, the U.S. military developed several systems for guid ing bombs to the target, rather than simply relying on gravity. Those systems culminated in the development of the first generation of laserguided bombs, which, when used during the Linebacker series of air operations in 1972, produced spectacular results and led to the term “surgical strike”—clearly a capability critical to discrete and discriminate use of force.³⁰

The U.S. military continued to develop precision‐​guided munitions after Vietnam, though the impetus for their development faded somewhat. In addition to bombs, precision‐​guided missiles with long range, such as the Tomahawk cruise missile, were developed during that period.³¹ The United States also developed cheaper precision weapons based on Global Positioning System (GPS) technology.³²

Precision‐​strike capability has been a major enabler of U.S. uses of force since the 1990s, from Bosnia and Kosovo to strikes attempting to kill Osama bin Laden. One of the limitations that emerged, particularly in the bin Laden case, was the problem of locating targets, especially mobile targets. In response, the United States initiated work on new systems for intelligence, surveillance, and reconnaissance, most notably those carried by unmanned drone aircraft. Such aircraft could relay images or signals intelligence back to target planners, who would then be able to direct precision strikes much more effectively.³³

Following the 9/11 attacks, both drones and precision strikes were a major element of the U.S. use of force. Precision strike capability was vital to the success of the 2001 campaign against al Qaeda and the Taliban in Afghanistan, where the United States had few troops.³⁴ Drones were armed with precision weapons as well, resulting in the ability both to find and to destroy targets in a discrete fashion without risking U.S. personnel. Drones also enabled a certain level of plausible deniabil ity for use in covert campaigns, such as targeting al Qaeda and other militants in Pakistan’s tribal regions.³⁵ That approach meant the United States could project discrete and discriminate force well short of a fullscale air campaign, much less an invasion.

The final change in the nature of U.S. military power is the creation of extensive and highly capable special operations forces. Although special operations forces are not a new concept, modern U.S. special operations forces are almost unique in size alone. The total personnel assigned to U.S. special operations is roughly two‐​thirds the size of the entire British army.³⁶ That quantity has a quality all its own, but in addition, the resources at the disposal of those forces for training and equipment are also unprecedented, producing a uniquely capable force with global reach.

That capability did not emerge from nowhere. As noted earlier, the failed rescue of the hostages in Iran in 1980 was an embarrassment to the United States. However, like the failures in bombing during Rolling Thunder, it also provided the impetus for major reform and improvement. The commission to investigate that failure, headed by retiredAdmiral James Holloway, argued for the creation of a multiservice (joint) special operations task force in order to provide the United States with a standing, integrated capability to conduct special operations, specifically with a focus on counterterrorism.³⁷

The Carter administration accepted the commission’s recommendation and established U.S. Joint Special Operations Command (JSOC)in 1980. It provided a framework for integrating some existing special operations units as well as newly created units. According to some public accounts, since its inception, JSOC has conducted operations ranging from hostage rescue to the capture or killing of warlords and war criminals.

In 1987, still unsatisfied with the state of U.S. special operations, Congress enacted legislation that created a four‐​star senior command over all special operations forces. U.S. Special Operations Command (USSOCOM, or SOCOM) created a high‐​level advocate for all special operations, not just for those under the authority of JSOC. SOCOM became a powerful institutional champion for special operations in the United States. That champion, combined with the general support given to U.S. special operations in both Congress and the executive branch, revitalized U.S. special operations capabilities beyond JSOC. Most notably, forces from JSOC conducted a daring raid deep into Pakistan on May 1, 2011, to kill Osama bin Laden.³⁸ That raid was perhaps the ultimate demonstration of the discrete and discriminate use of force.

Those three changes in the nature of U.S. military power are also mutually reinforcing. The shift from a military based on conscription to one based on professional volunteers has helped enable the training necessary to make full use of precision weapons while giving a highly qualified pool of personnel for special operations selection. Precision weapons provide effective fire support to small teams of special operators. Those teams, in turn, can provide intelligence and targeting for precision firepower.

Together, the changes represent a revolutionary change in the United States’ ability to use force. No longer are policymakers presented with the dilemma faced by the president and the public in the late 1970s and early 1980s of doing nothing or committing a major military force. Instead, individual terrorists or other suspects can be targeted almost globally by a combination of modern technology and a vast array of special operations forces.

The foregoing is in some ways a very positive development for dealing with the threat posed by substate actors. The ability to employ discrete and discriminate force in a wide variety of circumstances is extraordinarily useful for a calibrated response to those threats. Rather than invading a country or indulging in expensive and uncertain efforts to conduct nation building around the world, policymakers have an option that produces minimal domestic outcry (at least relative to the wars in Iraq and Afghanistan, to say nothing of Vietnam), that is not terribly expensive (again, at least in a relative sense), and that is often generally effective.

Those capabilities also offset to some degree the increase in the ability of substate actors to impose costs on the United States abroad. By limiting the exposure of U.S. personnel to “contested zones” to small numbers of elite personnel, those capabilities equally limit the impact of modern infantry weapons, IEDs, and suicide bombings. Thus, the increase in threat has been largely balanced by the increase in U.S. capability.

That balance should all be to the good, but the costs of using force have been reduced to such a low level that it has become an almost knee‐​jerk response to substate actors. Whether terrorists, rebels, or bandits, those actors all seem susceptible to the discrete and discriminate use of force, so why not exercise the option, which is almost costless? That has been the seductive effect of such developments: the use of force has swung from being prohibitively expensive (in political capital and risk) to almost free.³⁹

In addition to being almost free, use of force in that manner is perceived to be eminently controllable. U.S. political leaders don’t need to rely on troublesome allies or proxies to combat substate actors. Nor do those options require maintaining substantial knowledge of remote and inhospitable locations as much of the targeting of those operations can be collected from technical means, such as overhead imagery and signals intelligence. Thus, policymakers perceive the use of force to be cheap, controllable, and clean.

Yet that perception comes with a variety of problems. First, the idea that the use of force is cost free is true only if one neglects resulting externalities from that use of force. One prominent example of an externality that is much discussed, if not well researched, is the effect of drone strikes on recruiting for substate actors. Some have argued that drone strikes, from either the effect of collateral damage or the perception that U.S. use of force is illegitimate and murderous, or a combination, produce more recruits for terrorist and rebel organizations than they eliminate.⁴⁰ Thus, far from being an “easy” solution to the problem, the use of even discrete and discriminate force can be counterproductive. Although no reliable empirical data prove or disprove that alleged externality, it highlights the need to think beyond the obvious costs in evaluating the use of force.

Another possible externality is that the costs of the use of force will not remain low from a domestic or international politics perspective. At some point, the repeated use of force, even if that use of force does not directly affect the majority of Americans, may cause the public to tire of it. That result would likely happen slowly, if at all, but recent polls on the use of force in Syria suggest that the American public is weary of even discrete and discriminate use of force against state adversaries— weariness of the use of force against substate actors may not be far behind.⁴¹ Similarly, international support, which was never as robust as domestic support for such operations, may also decline further.

There are already two specific areas in which the discrete and discriminate use of force is increasingly contentious in the United States:

(a) the treatment of prisoners resulting from those operations, and

(b) the conduct of those operations against U.S. citizens. The first issue emerged in the immediate aftermath of the 2001 conflict in Afghanistan, as the United States captured terrorists and rebels there and elsewhere. Those combatants were not standard prisoners of war, as they were not fighting on behalf of a recognized state belligerent and, therefore, could not be held for the duration of declared hostilities. Yet trial under U.S. laws was problematic, either because the combatants had broken no U.S. laws or because information about their role in terrorist activity was classified and, therefore, inadmissible in a typical courtroom.

The U.S. government’s response was to hold those detainees as “unlawful enemy combatants” in various sites outside the United States, including the U.S. military base at Guantanamo Bay, along with secret facilities run by the Central Intelligence Agency (CIA) in various countries. However, that practice provoked a response by advocates of civil liberties, who noted that those individuals should be accorded some sort of due process to ascertain their guilt or innocence and should receive a defined sentence rather than indefinite detention without recourse. The Bush administration attempted to establish U.S. military tribunals to answer those complaints, but to date, such tribunals have been essentially ineffective in adjudicating cases.⁴²

President Obama entered the White House with the intention of closing Guantanamo within a year, yet he has thus far proved no more capable of resolving the central dilemma than his predecessor. The Obama administration’s attempts to move some detainees into the federal court system were widely criticized, and even his attorney general Eric Holder was forced to acknowledge that a “not guilty” verdict in federal court would essentially not be honored in the caseof some detainees. In testimony to the Senate Judiciary Committee, Holder replied to a question about the possible release of 9/11 plotter Khalid Sheikh Mohammed by noting, “We have taken the view that the judiciary does not have the ability necessarily to certainly require us, with regard to people held overseas, to release them.”⁴³ The detainee dilemma appears no closer to resolution now than it was a decade ago.

The result is that some argue that there is an increasing reluctance for the United States to take prisoners during operations against substate actors. Instead, to use the colloquial term, the policy has moved from “wanted dead or alive” to “wanted dead.” Although that may be an overstatement because some recent operations, such as the capture of Abu Anas al‐​Libi in Libya in early October 2013, suggest continued willingness to take prisoners, there has been a marked increase in the use of lethal operations over the past several years.⁴⁴

The second major issue is the use of force against American citizens. In 2001, U.S. and allied forces captured John Walker Lindh, the so‐​called American Taliban in Afghanistan. Since 9/11, Lindh was the first of several U.S. citizens linked to overseas terrorist activity directed against the United States. Yet such cases remained relatively rare, and the U.S. citizens involved were frequently apprehended and could be treated as criminals.⁴⁵ The United States was not forced to confront the possibility of targeting U.S. citizens with deliberate lethal force without trial as the consensus supporting the use of discrete, limited force against nonstate actors developed.

That approach changed with the case of Anwar al‐​Awlaki, a U.S. born Muslim cleric who became affiliated with al Qaeda and moved to Yemen. His sermons and exhortations on behalf of al Qaeda became prominent because of his charisma and command of colloquial English. Awlaki also apparently began to move beyond exhortation to active aid and planning of terrorist attacks.⁴⁶

The capacity of either Yemeni security forces or U.S. forces operating in Yemen to capture Awlaki was limited, because he was hiding in the almost ungoverned tribal hinterland. U.S. policymakers thus had to consider the possibility of targeting Awlaki in the same fashion as they targeted other terrorist leaders. After some debate between attorneys in the executive branch, the Justice Department’s Office of Legal Counsel produced a memorandum laying out a legal case for targeting Awlaki. He was killed in a drone strike in Yemen in September 2011.⁴⁷

Although the White House has argued that Awlaki was essentially a unique situation that did not set precedent for future use of force, it seems likely that at some future point, a similar situation will arise, and the issue will be revisited, with Awlaki serving as a precedent. However, there is at least some disquiet over the potential for targeting American citizens without trial. After all, intelligence agencies are legally required to obtain permission to monitor U.S. citizens’ communications from a special national security court (the “FISA court” named for the Foreign Intelligence Surveillance Act that created it). Thus, the strange circumstance was created whereby the U.S. government was, in all likelihood, required to go to court to monitor Awlaki’s cell phone but not to kill him.

Moreover, the issue of detainees and the killing of American citizens may be reinforcing each other in ways that may further undermine domestic support. If taking prisoners is not an option because of the detainee issue, then it follows that the use of lethal force against American citizens will become more likely. That consequence may limit support for any such operations.

The perception that the discrete and discriminate use of force is clean, in the sense of not requiring deep connections and human intelligence in remote and inhospitable regions, is also incorrect or at least incomplete. Although it is true that a substantial amount of targeting can be done from overhead and signals intelligence (SIGINT) platforms, human intelligence (HUMINT) is often just as important, if not more so. SIGINT offers global capability, but the volume of communications is simply too great to monitor simultaneously, so some sort of cue to which communications are important to monitor is needed. HUMINT often provides that initial cue.⁴⁸

HUMINT is particularly important against organizations that practice stringent operational security (OPSEC), such as avoiding the use of electronic communications in favor of face‐​to‐​face meetings or human couriers. Such OPSEC measures make coordination more difficult for clandestine organizations, asanyone who triesto arrange a meeting without the use of telephone, e‑mail, or social media can attest, but also make them secure. The hunt for Osama bin Laden, for example, was very long, in part because bin Laden did not use direct electronic communication, instead relying on couriers.⁴⁹

However, the detainee issue again raises its ugly head here. A major source of HUMINT is detainee interrogation, but if there is an unwillingness to detain, then targeting for the use of force becomes more difficult. Such operations may increasingly be self‐​defeating if in using force, the United States eliminates potential sources of future targeting intelligence while simultaneously undermining domestic support for such operations.

A further complication in reliance on the use of force is that it may be more effective against some substate actors than others. That is the reason conceptual clarity is necessary for understanding the natureof substate actors. Terrorists seeking to coerce the United States from remote locations face a variety of hurdles that the threat of discrete and discriminate use of force greatly complicates. For example, the need for OPSEC, as noted, can make coordinating transnational operations extremely difficult.

Likewise, terrorist organizations may have only a limited number of personnel who might be able to organize or conduct attacks on U.S. soil given the hurdles of getting into the country. The same may be true of individuals capable of financing operations or conducting effective recruitment. Eliminating even relatively small numbers of personnel may greatly hamper terrorist effectiveness, even if destroying the organization completely remains elusive.

In contrast, rebel groups may be substantially less vulnerable to the discrete and discriminate use of force. Because rebels are concerned about changing political order in their own region, power need not be projected over great distances, which makes OPSEC requirements potentially less onerous. As an example, contrast trying to arrange a suicide attack in New York from Karachi without using a phone or e‑mail with trying to arrange a suicide attack five blocks from your house without using a phone or e‑mail. The former is nearly impossible to do using means such as couriers, whereas the latter is relatively trivial to do using those same means.

Similarly, rebels may have a much greater recruiting base than terrorists have. In contrast to terrorists, rebels have a variety of grievances and incentives that they can use to attract followers to fight in their own country. As the bar to conduct operations for rebels is lower because they are acting locally rather than globally, many of those recruits may be effective at their tasks. Thus, some rebel groups may be much less vulnerable to the effect of discrete and discriminate use of force than many terrorist groups.

Bandits, in contrast, might or might not be vulnerable to the discrete and discriminate use of force, but there are risks in using force against them. First, use of force has the potential to alter the motivation of bandits from economic to political. Rather than simply being out to make money, they may decide to work actively against the political order or to attempt to coerce the United States to stop taking action against them. In short, bandits could be transformed by the use of force into rebels or terrorists.

Second, bandits may be filling a critical economic role in some areas or in answering some demands. Use of force against specific bandits may simply open up opportunities for new bandits. The use of force against Pablo Escobar and the Medellín drug cartel in Colombia in the 1990s appears to have had that result as the Medellín cartel was quickly replaced by the Cali cartel.⁵⁰

Finally, habitual reliance on the option of discrete and discriminate use of force has led to opportunity costs as other habits and capabilities are underused. Two of the most important are covert action and unconventional warfare capabilities, and the habit of relying on allies. Covert action and unconventional warfare mean relying heavily on surrogate forces rather than direct U.S. use of force, essentially contracting with or manipulating substate actors to combat other substate actors. That capability has experienced a rebirth since 2001, but it is still underused relative to discrete and discriminate force. Another approach relies on state allies of the United States to conduct operations, whether through the use of force or other means, to mitigate the threat of substate actors. Surrogates and allies provide capabilities that U.S. assets lack for combating substate actors. One example is the ability to take prisoners, which is complicated for the United States but may be substantially less so for surrogates and allies. For example, Jordan has been able not only to take prisoners but also, in some cases, to recruit them to work as double agents. Although one high‐​profile case backfired spectacularly when the double agent turned out to be a triple agent who then conducted a suicide bombing attack on a CIA base in Afghanistan, there are likely to be other examples that were more successful and, therefore, have not been publicly revealed.⁵¹

Surrogates may have similar advantages. In the 1990s, the CIA used surrogate forces to track Osama bin Laden and had plans to use that force to kidnap him. Though those plans were never approved, they point to possible opportunities today. For example, the recent operation to capture Abu Anas al‐​Libi, which provoked a major backlash against the United States, could have been conducted by a Libyan surrogate force rather than U.S. personnel. That alternative would have provided the United States with plausible deniability, given the overall chaos in Libya, yet it would have achieved the same effect.⁵²

Allies and surrogates also have a more persistent presence in a region, which lends longevity and continuity to efforts against substate threats. Unlike U.S. forces, which rotate into a theater and out again fairly rapidly, surrogates and allies are, in a sense, there for the duration. Also, unlike the attention of U.S. policymakers, which may wander from crisis to crisis, surrogates and allies are more likely to stay focused on threats in their immediate neighborhood from which they cannot simply walk away.

The central drawback that both allies and surrogates have compared with U.S. use of force options is the lack of control. Although U.S. forces and assets are extraordinarily responsive to political orders, surrogates and allies inevitably have their own agendas. It may require significant orchestration and negotiation to get them to conduct operations that U.S. assets could conduct almost instantly. In an impatient Washington environment, it is unsurprising that such lack of control is a significant drawback.

Lack of control also means that policymakers assume additional political risk in using allies and surrogates. The potential for unsavory activity by allies and surrogates, such as torture of prisoners, is one form of political risk stemming from lack of control. Indeed, the concern that potential surrogates for U.S. interests in Syria may be unsavory appears to be one of the major reasons that the United States has not been more aggressive in supporting surrogates there.

Finally, surrogates and allies are not an absolute panacea. In Somalia, for example, the United States has sought to support its allies Ethiopia and Kenya in taking the lead in operations against substate actors. In 2011, those two countries launched a limited incursion targeting substate actors such as the Al‐​Shabab militia in southern Somalia. Although those operations have had some success, they have not ended the threat posed by Al‐​Shabab, and Kenya has suffered retaliatory attacks from Al‐​Shabab.⁵³

Threats from substate actors are real if limited, which necessitates a real if limited U.S. response. U.S. strategy must, as with any good strategy, balance the costs of action with the risks of inaction. That strategy requires a full appreciation of the costs of the discrete and discriminate use of force, which, although relatively low in U.S. physical and domestic political costs, has less obvious but equally important costs, such as the potential to fuel recruitment for substate threats or the risk of transforming one type of substate threat into another.

The foregoing is not an argument for the United States’ abandoning the discrete and discriminate use of force against substate threats. Instead, there must be a diversification of policy options to confront those threats. Use of force should be one tool among many, particularly the use of allies and surrogates, to combat such threats.

That approach will require a shift in U.S. defense and intelligence community investments, some of which are already quietly under way.

U.S. SOCOM commander Admiral William McRaven’s plan to builda global special operations force network is one example of such a reorientation regarding certain allies. A similar reorientation toward more willingness and capability to support surrogates would likewise be useful. The latter may also be under way already, but if so, it is understandably opaque.

In turn, that shift will also require greater patience and risk taking in Washington, two traits that are not in large supply inside the Beltway. Greater patience is required because those investments take time and continual effort to pay off, yet because of the relative lack of control over allies, that shift requires more assumption of risk. Yet the Obama administration seems to have concluded that the discrete and discriminate use of force may have reached the point of diminishing returns. We hope policymakers will realize that other ways to achieve the same ends may be possible.

Perhaps even more important is determining which, if any, substate actors actually constitute a threat. Drawing the line is critical but difficult in many cases. The guiding principle should be to focus on the essence of the substate actor and its actual goals, rather than lumping together all substate actors of a certain type. Every Islamic militant or narcotrafficker is not an equal threat to the United States.

In evaluating substate threats, there are two concerns. The overriding concern must be those threats that consistently seek to threaten the U.S. homeland. That concern obviously includes al Qaeda central and its Arabian Peninsula affiliate, yet it does not, at present, seem to include its Iraq and Syria affiliates. Likewise, it does not include the Haqqani network, which despite its lethality to coalition forces in Afghanistan has not been significantly associated with attacks or attempted attacks on the U.S. homeland.⁵⁴

The second concern is the threat to U.S. allies. Although that concern is important, it is secondary to threats to the homeland. Moreover, the U.S. response would likely be better served through long‐​term efforts to enhance the security capabilities of the ally rather than the discrete and discriminate force. Though that conclusion may seem obvious as noted, it can often be difficult in practice to estimate the effect of the use of force on substate threats. What seems to be a sensible use of force may have unanticipated costs.

Alleged U.S. use of force in Pakistan highlights the complicationsof the use of discrete and discriminate force. To bolster and placate Pakistan, which was confronting a severe threat by numerous Islamic militants, the United States allegedly expanded drone strikes to include leaders of the Pakistani Taliban in 2009. Before that expansion, the Pakistani Taliban had shown little interest in attacks on the U.S. homeland, instead focusing on regional clashes.

Yet after Pakistani Taliban leader Baitullah Mehsud was killed by an alleged drone strike in August 2009, the group has shown more interest in attacks on the continental United States. Most notably, the Pakistani Taliban provided training and support to failed Times Square bomber Faisal Shahzad.⁵⁵ At the same time, the Pakistani Taliban continue to be a threat to Pakistan, despite the U.S. use of force.⁵⁶ Thus, the U.S. use of discrete and discriminate force changed the nature of the Pakistani Taliban threat without substantially aiding Pakistan—hardly an approach to avoid “vexation and exhaustion.”

To be fair, the use of force to support Pakistan against the Pakistani Taliban may have still been on net a positive for U.S. security. It helped ensure continued cooperation, within limits, from the Pakistani government with a variety of issues, from drone strikes to supply lines into Afghanistan. But a fuller appreciation of the potential costs and limitations of the use of force may have changed U.S. expectations for the “return on investment” of targeting Pakistani Taliban leaders.

Finally, the expectations for what any set of policy options can be reasonably expected to achieve need to be better calibrated. The discussion in recent years of whether al Qaeda has been defeated has missed the point. Even if al Qaeda has been defeated, other substate threats will emerge, whether terrorists and rebels in Mali and Syria or pirates in Somalia. Instead, policymakers must accept that substate threats will always be with us and must prepare to mitigate rather than eliminate the threats they pose. That will be the true management of savagery.