Center for a New American Security

Intelligence reform is once again in the air, and this time the bogeyman is the "militarization" of the Central Intelligence Agency (CIA). As Mark Safranski notes, there is something deeply bizarre about the idea that an organization created from the bones of a paramilitary covert action force (the Office of Strategic Services) and frequently involved in joint military ventures with special operations forces (like the Vietnam-era Phoenix Project) should be blamed for engaging in large-scale collaborative military ventures. The frequency with which observers call for the Agency to reject militarization and pursue the supposedly more pure activities of intelligence collection and analysis suggest a lack of historical knowledge of the CIA's paramilitary past.

The CIA was built to perform both covert action and intelligence collection. Aside from a brief period in the 1990s, the Agency has engaged in both. Creating and managing paramilitary armies, running off-the-books air forces and engaging in political action and influence campaigns has been the Agency's bread and butter from Southeast Asia to Latin America and plenty of places in between. The military and other elements of the interagency has been a natural partner for the CIA, particularly in waging sustained and violent campaigns against sub-state actors. This is not to deny the importance of Sherman Kent and the Agency's strategic intelligence mission. But focusing on Kent alone makes us forget that for every Kent there was also a James Jesus Angleton or a Kermit Roosvelt. And for every Angleton and Roosevelt there was also a MACV-SOG operator or a deniable pilot bombing a Third World battlefield in the Cold War.

It was precisely the lack of good paramilitary options during the 1990s that spawned policymaker demand for unmanned aerial vehicle strikes and greater interagency fusion between the CIA and the Joint Special Operations Command (JSOC). Note too that for better or worse, the Agency and its military counterparts have also returned to form in Central Asia and Africa with local paramilitary proxy forces and secret (non-UAV) air forces in geopolitical hot spots. Merely laying out the continuum of military/intelligence fusion in past and present eras of covert activity isn't enough. What is needed is a framework for understanding why the Agency and the military remain so intricately tied together, and why paramilitary missions are likely to continue in some form no matter what is said about the benefits of rejecting targeting and returning to intelligence collection.

This framework is absent from 90% of conversations about military and intelligence fusion, and the debate needs better perspective. Let's start with the purpose of national power.  There are many sides of power, defined at a minimum as the ability to shape the ability of actors to determine their circumstances and fate. Some aspects of power are inherently compulsory in nature, meaning they hinge on the capability to force someone (an individual) or something (an social entity) to do what you want. Others have to do with institutional heft or producing a ideology capable of mobilization.

However, the intermediary step between the capability, structure, or discourse in question and the resulting outcome is often some form of control. In counterinsurgency campaigns, it is difficult to build states without the ability to prevent the enemy from interfering. Creating such control requires violence when the violent objector to the policy objective remains on the field. More broadly, states have developed elaborate means of creating control over domestic and international social environments. As Erin Simpson reminds us, even architecture can help bolster state control over a restive polity through design that facilitates tactical advantage and movement by local security forces.

One means of control is information. Counterinsurgency theorists have often emphasized the importance of cartography, demographic information, relational data, and superior means of information processing because greater understanding an operational environment lowers the costs of control. The anarchist social scientist James C. Scott alleged (with a fair degree of infamy) that certain hill tribes in Asia resisted developing written languages in order to avoid incorporation into centralized polities. Counterintelligence and deception operations deny and misdirect adversaries in search of information. The phrase "information is power" is banal but there is some use for it. 

Intelligence analysis takes information and data and transforms it into actionable products for policymakers. Unsurprisingly, intelligence has always been tied to control, domestically or internationally. Strategic indicators and warning (I&W) helps deny control to a foreign enemy by providing advance warning of attack. The role of domestic intelligence in maintaining domestic political control is obvious. Intelligence does not guarantee control but it is certanly part of the requirements for generating control in a social environment.

Control can also be gained through organized social action, which can directly deny, disarm, destroy, or otherwise thwart an objector to state policy. Organized violence is a form of social action, but so can be a political mobilization, construction of an alternative institution, or calculated erosion of a target social structure or entity. Many covert operations use combinations of violence, hidden influence, erosion, and mobilization to achieve control. There's a long CIA history of militarizing ethno-religious groups that are on the bottom of a given country's power structure or arming insurgents and providing them with military support.

So what does all of this have to do with intelligence? Plenty. Policymakers did not develop intelligence agencies to do "intelligence," a term that only represents a fraction of what many intelligence organizations across the world do. Let's not confuse a given institutional design, which can shift radically over time, with core purpose. Policymakers have always wanted intelligence agencies, intelligence entrepreneurs, or specialized ad hoc groupings to improve their ability to generate control in the following ways: collection of information, manipulation, and direct coercive action, often in concert with other agencies. Special operations historian Simon Anglim lays out how covert operations fit in:

There are grey areas in which states clash short of open warfare when use of subversion, sabotage and fighting by local proxies may be a preferred strategic option to overt commitment of regular forces; moreover, given that much non-violent covert activity aims at undermining the target state’s military preparedness and will to fight, and steering its strategic decision-making processes, there are important strategic dimensions here, also. A covert operation, therefore, is a single mission aimed at creating a particular situation in another country with concealed means and intent. Non-violent covert operations create disaffection among the target state’s population, weakening its will to affect the world around it, or steer surreptitiously its decision-making via placing agents in key positions. Violent covert operations include sabotage, assassination, and paramilitary support of armed insurgency against the opposing power.

Many covert activities happen during wartime and in the context of larger military campaigns. During the Pancho Villa expedition Pershing attempted to use secret agents to poison Pancho Villa outright. The infamous Force Research Unit (FRU) used agents of intelligence to further British strategy in the struggle against the Provisional Irish Republican Army. The CIA carried out extensive covert operations of a paramilitary nature in the Korean War and did so in cooperation with special operations soldiers as part of MACV-SOG in the Vietnam war. And there was the OSS in WWII, the proliferation of specialized British military intelligence and special operations groups, and the operation of private air forces before US entry into the war.

Policymakers want a variety of means to realize control. When regular military means are too blunt a tool and diplomacy too soft, they will clamor for what they (rightly or wrongly) consider to be a subtle knife. In Ghost Wars, Steve Coll noted that every time the Pentagon could not give the President actionable options for raids in pre-9/11 Afghanistan without rendering the plans politically useless. In the eleven years since 9/11, successive Presidents have clearly rejected such a calculus and demanded that the CIA and the Pentagon work together to produce covert options for pursuing enemies of the state. In particular, the line between the CIA has always been fairly porous.  As SOF scholar Nick Prime often noted on Twitter, military and CIA often combine to realize common missions, exchange personnel, and closely collaborate.

Finally, like any good bureaucratic actor the CIA jockeys to serve a pressing need. The most pressing need today, as perceived by policymakers, is the war effort. Militarization goes beyond just adapting to the demands of policymakers, as a former classmate pointed out to me. If the President did not want a military-focused CIA, would he have appointed a former general to head it? And David Petraeus, unlike fellow General Michael Hayden, was not dual-hatted.

If the CIA decides to go "back to intelligence," policymakers will rightly conclude that the Agency is only doing half of its job. It seems paradoxical to bemoan the militarization of intelligence when the likely outcome of muzzling the CIA's paramilitary organs will be greater Pentagon conduct of covert affairs. As Robert Caruso once explained, it is not clear that this structure will result in greater transparency. DoD has many more ways than the Agency to frustrate such inquiries. A CIA weak on paramilitary-focused covert action will guarantee a more militarized intelligence structure than the Intelligence Community we have today.

Yet there is some merit to criticism of the CIA's post-9/11 focus. The emphasis on the targeting cycle has come at the expense of strategic intelligence. If intelligence agencies cannot deliver deliver credible assessments of intelligence useful to the formation of strategy, they are also doing half their jobs. The strategic effect derived from post-9/11 covert action has also been mixed. Data problems make it hard to assess the impact of the air war in Pakistan, and in Yemen the covert campaign has been arguably counterproductive. There are reasonable concerns about the long-term impact of the CIA losing its significant institutional advantages in complex counterintelligence, non-targeting covert operations, and intelligence collection.

But blaming paramilitary covert operations alone ignores the fact that covert operations in war serve strategy. And when policy and strategy have an overly military character, it is not surprising that myopia will set in as intelligence efforts are shifted to fit operational demands. Covert operations in Vietnam and its neighbors were supposed to support an overarching war effort that in and of itself was horribly flawed. In Yemen the United States is the counterinsurgency force for the Yemeni government, and the failures of targeted killings must be considered within the context of Washington's poor conduct of the overall campaign.

The focus on the targeting cycle owes itself to policymaker demand. Listening to some critiques of CIA/JSOC operations for militarization of intelligence, one would never realize that the nation has been at war for eleven years. If the CIA did not radically shift to support operational military demands it would have likely endured the same tongue-lashing Robert Gates gave the Air Force for not supporting the COIN mission. Nearly every element of the United States government was told to support the wars, and the CIA was no exception to the general rule. During times of war or situations in which the chance of war is high, policymakers tend to want the military and the CIA to cooperate to advance their strategic aims or otherwise help realize policy. if you want a less militarized IC, you ought to demand a policies less reliant on violent and/or coercive means. And in particular, policies that demand an expansive political-military effort across a large portion of Africa, the Middle East, and Central Asia.

Finally, critics engrossed in the ethical quandries of direct action should not ignore the significant moral compromises inherent in day-to-day intelligence collection (particularly human intelligence operations). There is a reason why Americans have always been traditionally uncomfortable with intelligence agencies, and we have never quite given up the basic sentiment that "gentlemen don't read other gentlemen's mail." Intelligence collection has often threatened civil liberties, and convincing vulnerable informants to betray their country, group, or tribe is fraught with ethical perils. Today, the ethical challenge of human intelligence operations are often forgotten. But they were an omnipresent problem during the Cold War, and one recognized in even fictional literature on intelligence like John LeCarre's novels.

So what to be done? First, any process of intelligence reform has to incorporate what policymakers fundamentally desire out of organizations like the CIA. In an ever more open-ended struggle against al-Qaeda and affiliated groups, this will be covert action in cooperation with the military. The difficult challenge lies in resolving the Title 10/50 conflict and developing greater transparency while still preserving institutional relationships and tactics, techniques, and procedures built up during the post-9/11 wars. The CIA will also have to balance such demands with core intelligence collection and assessment. The frustrating part about this is that the CIA's ability to pursue better intelligence collection and analysis is inherently constrained by policymaker demand. Another Afghanistan and we will be likely to see much more attention and resources focused on targeting. The fictional CIA honcho in Zero Dark Thirty yells "give me targets!" because he likely has someone above the Agency also breathing down his neck.

Most importantly, the United States must also understand that covert operations are also governed by policy and strategy as well. As Micah Zenko's research shows, "discrete military operations" are effective in the context of an overarching strategy. And it goes without saying that covert operations cannot rescue a bad policy. Unfortunately, much of the history of covert operations in America is often precisely that--failed attempts to rescue bad policies with spooks and door-kickers. Covert operations have their limits and it would behoove us to spend more time trying to understand their inherent constraints. But setting up a false binary between a paramilitary and intelligence CIA won't help us do that.

The controversy of the American targeted-killing program, and especially the resurgence of covert paramilitary and military action, has inspired a great deal of concern about the accountability and oversight of America’s supposed new ways of war. Does the lack of risk they offer encourage the Congress, media, and public to stay silent? One of the most prominent scholars of military robotics, P.W. Singer, recently put out an article that reiterated an argument he makes about the decline in the accountability of American wars, as exemplified in the drone program:

In democracies, there have always been deep bonds between the public and its wars. Citizens have historically participated in decisions to take military action, through their elected representatives, helping to ensure broad support for wars and a willingness to share the costs, both human and economic, of enduring them.

In the U.S., our Constitution explicitly divided the president’s role as commander-in-chief in war from Congress’s role in declaring war. Yet, these links and this division of labour are now under siege as a result of a technology that our founding fathers never could have imagined.

We don’t have a draft anymore. Less than 0.5 per cent of Americans over 18 serve in the active-duty military. We do not declare war anymore. The last time Congress actually did so was in 1942 – against Bulgaria, Hungary, and Romania. We don’t buy war bonds or pay war taxes anymore. During the Second World War, 85 million Americans purchased war bonds that brought the government $185 billion. In the last decade, we bought none and instead gave the richest five per cent of Americans a tax break.

And now we possess a technology that removes the last political barriers to war. The strongest appeal of unmanned systems is that we don’t have to send someone’s son or daughter into harm’s way. But when politicians can avoid the political consequences of the condolence letter – and the impact that military casualties have on voters and on the news media – they no longer treat the previously weighty matters of war and peace the same way.

For the first 200 years of American democracy, engaging in combat and bearing risk – both personal and political – went hand in hand. In the age of drones, that is no longer the case.

This narrative exemplifies a civil mythology under final assault from the robotic barbarians at the gates. Unfortunately, history itself tells a far messier story. For one thing, the notion that there are always deep bonds between the public and the war-fighting effort is false. I have tackled the question of the draft previously on this blog, but the rest of the arguments merit further scrutiny.

For one, the Constitution’s demands on Congressional oversight in war have never been so clear, nor so linear in their erosion. The U.S. fought several wars without a formal declaration – and even without direct Congressional authorization – before it ever formally declared war in 1812. In some cases, such as the Quasi-War and the Barbary Wars, these were authorized by Congressional statutes short of a formal declaration. In 1801, Congress passed the Naval Peace Establishment Act, and Jefferson cited Congress’s funding of the military capacity as sufficient authorization for its use against hostile powers. A State Department directive told the U.S. Navy that if the Barbary states declared war on the U.S., then the Navy was to “protect our commerce & chastise their insolence – by sinking, burning, or destroying their ships & Vessels wherever you shall find them.”

Of course, Jefferson was hesitant to expand this further than defense and limited retaliation, but even he did not believe a formal declaration, nor, obviously, any kind of conscription, was necessary for waging offensive war. What he received was a series of Congressional statutes expanding the fleet and specifically authorizing expanded military action against the Barbary states.

The Indian wars were justified on much the same logic. At no point did the U.S. formally declare war against the Indians. By the period of the Seminole Wars it was well-established that Congress recognizing hostilities and appropriating resources to the combat established constitutional recognition of a conflict. Insofar as Congress receives statutory notification and continues to defray the costs of conflict, it legitimizes war as constitutional. The differences between a Congressional authorization for using force and a formal declaration are statutorily meaningful, but both are legitimate with respect to the Constitution.

The deep civic bonds have actually generally been quite shallow. State militias were called up in local wars for military geographical reasons, but the burning of the capital in 1814 failed to merit a draft. The AUMF, NDAA, and War Powers Resolution all constitute a system of Congressional compliance to Presidential military initiative, in which war is retroactively legitimized through post hoc defrayment.

The U.S. Navy, with its peacetime establishment and broad writ to conduct “small wars” and punitive expeditions (as well as a Marine Corps with similar advantages), did far more to undermine the political barriers to U.S. wars than drones have or likely will. Expeditionary warfare by forces inherently limited in their political costs of extraction is as old as the republic itself.

The very concept of covert action helped too, and the idea of a secret air force predates the CIA itself. Roosevelt’s Flying Tigers, approved before U.S. entry into WWII using government money laundered through a contractor and lend-lease, sought to secretly put dozens of aircraft into China to fight Japan. Manned aircraft, along with a PMC, and an extralegal or illegal authorization by a frustrated executive began what was planned to be a covert war. December 7, 1941, not deep civic responsibility, saved it from being remembered as such. Later, the CIA was flying secret air forces for Cubans and Congolese. WWII era aircraft flew secret wars, but covert action itself was the real mechanism for reducing political costs. That it now happens to employ robots rather than deniable pilots, foreign mercenaries, or nameless spooks allows changes in quantity more than essential quality.

Yet many insist that drones, by removing the threat of casualties, undermine oversight and accountability because politicians can avoid legislative backlash and media scrutiny. Even recent history does not bear this out.

For example, at least 17 Americans died in Operation Enduring Freedom-Philippines, including some in militant attacks and not simply accidents or other causes. Operation Enduring Freedom-Horn of Africa has its casualties too. The media noted their deaths but there was no backlash. Bemoaning the lack of media coverage of Afghanistan – Afghanistan! – is a cliché of war commentary that will be decade old before the drawdown. Relative to the number of U.S. personnel committed, I would wager the targeted killing campaign is far better covered than Afghanistan today, and it is certainly out of proportion in terms of the casualties that the personnel supporting the war suffer.

Indeed, the wars in Iraq and Afghanistan themselves do not suggest that the link between perpetual war and lack of risk is so straightforward. What, exactly, has Congressional oversight there saved us from? For all the bemoaning of the “drone wars,” they have broad public approval, incredibly little Congressional criticism – by the cost-defrayment standard, continual support – and can we really say their consequences are so much more deleterious to the body politic and the public trust than the disaster that was the choice to invade Iraq? To ignore, escalate, and bungle in Afghanistan?

We can’t blame the drones for the U.S. war in Yemen, where US SOF and clandestine agents watched from the ground when 2002’s lethal drone strike came down, or where cruise missiles and most likely F-15Es take part in the bombardment. We can’t blame the drones for the U.S. war in Somalia, where naval guns, AC-130s and helicopters, along with JSOC, operated for years before the Predators and Reapers let loose missiles there. In any case, neither of these states really have the air defense capability, or the intention, to challenge U.S. airpower. Are we really to believe the risk of a plane crash is why policymakers switched to drone strikes?

As for Pakistan, the model of accountability that Singer holds up, the bin Laden raid, involved a lot of deliberation and careful consideration, to be sure – but it was done entirely in secret. That we even know of its deliberations so intimately is because, for basically everyone involved, it’s a good story. That we use drones there and not conventional aircraft is not because of American casualty aversion, but because it is what the Pakistani government appears to accept – and these strikes frequently cease or slacken when Pakistani and U.S. relations come too close to the brink. Political costs retain veto power, but in covert action, they are quiet and indirect.

The fault lies not in our drones, but in ourselves. The reason our wars – secret or no – are so poorly managed are because of the policy process itself and the goals it seeks, alongside the incredible capability of the U.S. military and federal government which lets them sustain the weight and persevere through so many missteps and failures. The draft does not stop failing wars, overt or covert, as we learned from Vietnam and the “secret wars” surrounding it. That the condolence letter of a pilot crashing his aircraft in Yemen might be the difference between peace and war seems proper, but what would make its power so much greater than those for the advisors and the spotters, or the vastly larger number of letters for the fallen of Afghanistan, which was sickeningly, but unsurprisingly, absent from the general election? The political silences that enable these processes are older than we care to admit. It is not just that we cannot turn back time, but that there is no extended length of time much better to turn back to. Before drones were, these kinds of wars were there, waiting for them.