Three Pillars of Counterinsurgency
Dr David J. Kilcullen*
Remarks delivered at the U.S. Government Counterinsurgency Conference,
Washington D.C., 28 September 2006
Introduction
We meet today in the shadow of continuing counterinsurgencies that have cost thousands
of lives and a fortune in financial, moral and political capital. And we meet under the
threat of similar insurgencies to come. Any smart future enemy will likely sidestep our
unprecedented superiority in traditional, force-on-force, state-on-state warfare. And so
insurgency, including terrorism, will be our enemies’ weapon of choice until we prove
we can master it.1
Like Bill Murray in Groundhog Day, we are going to live this day
over, and over, and over again — until we get it right.
So we seek a common doctrine to integrate national power against the threat. This has
happened before, it turns out.
The United States produced an inter-agency counterinsurgency doctrine in 1962. Called
the Overseas Internal Defense Policy (OIDP)2
, it was “prepared by an Interdepartmental
Committee consisting of representatives of State (Chair), DOD, JCS, USIA, CIA and
AID.”3
It was approved under National Security Action Memorandum 182 of 24 August
1962, signed by McGeorge Bundy4
and overseen by a Special Group (CounterInsurgency),
comprising “the Chairman of the Joint Chiefs of Staff, the Deputy Secretary
of Defense, the Director of Central Intelligence, the heads of AID and USIA, a staff
member of the National Security Council, and…the Attorney General of the United
States”.5
OIDP lays out a framework for whole-of-government counterinsurgency,
assigns responsibilities and resources, and explains what each agency brings to the fight.
Why the history lesson? Because last time we tried this, it did not work very well. OIDP
was classified, and while it informed senior leaders it filtered only fitfully down to the
field. It was applied in only the minor campaigns of the day. And it lasted only until
1966. As Vietnam escalated, OIDP (used during the advisory phase of the war) was
dropped and the campaign was handed off to the conventional military and the State
Department’s “A” Team of Europeanists and Cold Warriors.6
And so, as many have
observed, our problem is not that we lack doctrine but that we continually forget, relearn,
discard our corporate knowledge, and treat as exceptional one of the most common
forms of warfare.7
Today, things are even more complicated than in 1962. To be effective, we must marshal
not only all agencies of the USG (and there are more than 17 agencies in the foreign
policy arena alone8
), but also all agencies of a host nation, multiple foreign allies and
coalition partners, international institutions, non-government organizations of many
*
Chief Strategist, Office of the Coordinator for Counterterrorism, U.S. Department of State.
Correspondence address: 2201 C St NW Washington D.C. 20520 e-mail kilcullendj@state.gov This
presentation represents the author’s personal opinions only.
2
national and political flavors, international and local media, religious and community
groups, charities and businesses. Some have counterinsurgency doctrine that is more or
less compatible with ours. Some have different doctrines, or none. Some reject the very
notion of counterinsurgency — but all must collaborate if the conflict is to be resolved.
This means we need a way to generate purposeful collaboration between a host of actors
we do not control. No doctrinal handbook will ever be flexible enough for such a fluid
environment (though, something tells me, we will develop one anyway). Rather, we need
an easily grasped mental model that helps individuals and agencies cooperate, creates
platforms for collaboration, and forms a basis for improvisation. In conventional war we
might call this an “operational design”, or “commander’s intent”. I will call it a “model”.
There are two parts to this model. The first is a description of the “conflict ecosystem”
that forms the environment for 21st century counterinsurgency operations. The second is
a tentative framework for whole-of-government counterinsurgency in that environment.
The Conflict Environment
An insurgency is a struggle for control over a contested political space, between a state
(or group of states or occupying powers), and one or more popularly based, non-state
challengers.9
Insurgencies are popular uprisings that grow from, and are conducted
through pre-existing social networks (village, tribe, family, neighborhood, political or
religious party) and exist in a complex social, informational and physical environment.10
Think of this environment as a sort of “conflict ecosystem”.
It includes many independent but interlinked actors, each seeking to maximize their own
survivability and advantage in a chaotic, combative environment. Pursuing the ecological
metaphor, these actors are constantly evolving and adapting, some seeking a secure niche
while others seek to become “top predator” or scavenge on the environment. Some actors
existed in the environment before the conflict. They include government, ethnic, tribal,
clan or community groups, social classes, urban and rural populations, and economic and
political institutions. In normal times, these actors behave in a collaborative or
competitive way: but now, due to the internal power struggle, they are combative and
destructive. The relatively healthy competition and creative tension that sustains normal
society has spun out of control, and the conflict threatens to destroy the society.
This new state of the environment also produces new actors. These include local armed
organizations, and foreign armed groups drawn into the conflict from outside. Often, that
includes intervening counterinsurgent forces such as ourselves. Foreign terrorists are also
increasingly “swarming” from one conflict to another in pursuit of their global agenda. In
addition, the conflict produces refugees, displaced persons and sometimes mass
migration. It creates economic dislocation, leading to unemployment and crime, and
creating armed groups such as bandits, narcotics traffickers, smugglers, couriers and
black marketeers.
This might be illustrated graphically as in figure 1.
3
It is critically important to realize that we, the intervening counterinsurgent, are not
outside this ecosystem, looking in at a Petrie dish of unsavory microbes. Rather, we are
inside the system. The theater of operations is not a supine, inert medium on which we
practise our operational art. Rather it is a dynamic, living system that changes in response
to our actions and requires continuous balancing between competing requirements.
Where the counterinsurgent differs from other actors is largely a matter of intent. Like
other players, we seek to maximize our survivability and influence, and extend the space
which we control. But unlike some other players (the insurgents, for example) our intent
is to reduce the system’s destructive, combative elements and return it to its “normal”
state of competitive interaction. This has sometimes been expressed as “bringing
democracy” but, of course, democratic processes without the foundation of a robust civil
society may simply create instability and perpetuate conflict. Thus, whatever our political
objective, our functional objective is to impose a measure of control on the overall
environment. But in such a complex, multi-actor environment, “control” does not mean
imposing order through unquestioned dominance, so much as achieving collaboration
towards a set of shared objectives.
If this sounds soft, non-lethal and non-confrontational, it is not: this is a life-and-death
competition in which the loser is marginalized, starved of support and ultimately
destroyed. The actors mount a lethal struggle to control the population. There is no
known way of doing counterinsurgency without inflicting casualties on the enemy: there
is always a lot of killing, one way or another. But killing the enemy is not the sole
objective — and in a counterinsurgency environment, operating amongst the people,
force is always attended by collateral damage, alienated populations, feuds and other
unintended consequences. Politically, the more force you have to use, the worse the
campaign is going. Marginalizing and out-competing a range of challengers, to achieve
control over the overall socio-political space in which the conflict occurs, is the true aim.
4
Remembering that this is simply a theoretical model, and thus a brutal oversimplification
of an infinitely complex reality, how might we seek to operate in this environment?
A framework for inter-agency counterinsurgency
Obviously enough, you cannot command what you do not control. Therefore, “unity of
command” (between agencies or among government and non-government actors) means
little in this environment. Instead, we need to create “unity of effort” at best, and
collaboration or deconfliction at least. This depends less on a shared command and
control hierarchy, and more on a shared diagnosis of the problem, platforms for
collaboration, information sharing and deconfliction. Each player must understand the
others’ strengths, weaknesses, capabilities and objectives, and inter-agency teams must
be structured for versatility (the ability to perform a wide variety of tasks) and agility (the
ability to transition rapidly and smoothly between tasks).
A possible framework for inter-agency counterinsurgency operations, as a means to
creating such a shared diagnosis, is the “three pillars” model depicted at Figure 2.
This is a framework, not a template. It helps people see where their efforts fit into a
campaign, rather than telling them what to do in a given situation. It provides a basis for
measuring progress and is an aid to collaboration rather than an operational plan. And
clearly, it applies not only to counterinsurgency but also to peace operations, Stabilization
and Reconstruction, and complex humanitarian emergencies. The model is structured as a
base (Information), three pillars (Security, Political and Economic) and a roof (Control).
This approach builds on “classical” counterinsurgency theory, but also incorporates best
practices that have emerged through experience in peacekeeping, development, fragile
states and complex emergencies in the past several decades.
5
Within this “three pillars” model, information is the basis for all other activities. This is
because perception is crucial in developing control and influence over population groups.
Substantive security, political and economic measures are critical but to be effective they
must rest upon, and integrate with a broader information strategy. Every action in
counterinsurgency sends a message; the purpose of the information campaign is to
consolidate and unify this message. It includes intelligence collection, analysis and
distribution, information operations,11
media operations (including public diplomacy) and
measures to counter insurgent motivation, sanctuary and ideology. It also includes efforts
to understand the environment through census data, public opinion polling, collection of
cultural and “human terrain” information in denied areas. And it involves understanding
the effects of our operations on the population, adversaries and the environment. Clearly,
not all actors will collaborate in these efforts; but until an information base is developed,
the other pillars of counterinsurgency cannot be effective. Importantly, the information
campaign has to be conducted at a global, regional and local level — because modern
insurgents draw upon global networks of sympathy, support, funding and recruitment.
Resting on this base are three pillars of equal importance. Indeed, as Figure 2 illustrates,
unless they are developed in parallel, the campaign becomes unbalanced: too much
economic assistance with inadequate security, for example, simply creates an array of
soft targets for the insurgents. Similarly, too much security assistance without political
consensus or governance simply creates more capable armed groups. In developing each
pillar, we measure progress by gauging effectiveness (capability and capacity) and
legitimacy (the degree to which the population accepts that government actions are in its
interest). This approach is familiar to anyone who has participated in a USAID conflict
assessment, or worked on fragile states or complex humanitarian emergencies. It has a
solid basis in empirical field experience in the aid and development community.12
The security pillar comprises military security (securing the population from attack or
intimidation by guerrillas, bandits, terrorists or other armed groups) and police security
(community policing, police intelligence or “Special Branch” activities, and paramilitary
police field forces). It also incorporates human security, building a framework of human
rights, civil institutions and individual protections, public safety (fire, ambulance,
sanitation, civil defense) and population security. This “pillar” most engages military
commanders’ attention, but of course military means are applied across the model, not
just in the security domain, while civilian activity is critically important in the security
pillar also. Clearly, also, security is not the basis for economic and political progress (as
some commanders and political leaders argue). Nor does security depend on political and
economic progress (as others assert). Rather, all three pillars must develop in parallel and
stay in balance, while being firmly based in an effective information campaign.
The political pillar focuses on mobilizing support. As for the other pillars, legitimacy and
effectiveness are the principal dimensions in which it is developed. It comprises efforts to
mobilize stakeholders in support of the government, marginalize insurgents and other
groups, extend governance and further the rule of law. A key element is the building of
institutional capacity in all agencies of government and non-government civil institutions,
and social re-integration efforts such as the disarming, demobilization and reintegration
(DDR)13
of combatants. Like the security pillar for military forces, the political pillar is
the principal arena for diplomatic and civil governance assistance efforts — although,
again, civil agencies play a significant role in the security and economic pillars also.
6
The economic pillar includes a near-term component of immediate humanitarian relief, as
well as longer-term programs for development assistance across a range of agricultural,
industrial and commercial activities. Assistance in effective resource and infrastructure
management, including construction of key infrastructure systems, is critically important.
And tailoring efforts to the society’s capacity to absorb spending, as well as efforts to
increase absorptive capacity, underpin other development activities.
These three pillars support the overarching objective of control, which — as we have
seen — is the counterinsurgent’s fundamental aim. The aim is not (as some have argued)
simply to create stability. Stability may actually not be our objective, as the President
emphasized in his recent speech to the United Nations General Assembly, when he
observed that “on 9/11, we realized that years of pursuing stability to promote peace left
us with neither. Instead, the lack of freedom made the Middle East an incubator for
terrorism. The pre-9/11 status quo was dangerous and unacceptable.”14
Moreover, even
if we do seek stability, we seek it as a means to an end, a step on the way to regaining
control over an out-of-control environment, rather than as an end in itself.
In achieving control, we typically seek to manage the tempo of activity, the level of
violence, and the degree of stability in the environment. The intent is not to reduce
violence to zero or to kill every insurgent, but rather to return the overall system to
normality — noting that “normality” in one society may look different from normality in
another. In each case, we seek not only to establish control, but also to consolidate that
control and then transfer it to permanent, effective and legitimate institutions.
Operationalizing the “Three Pillars”
If this model represents a possible framework for inter-agency counterinsurgency, how
might we apply it in practice? Arguably, the basis for doing so exists already, in National
Security Presidential Directive 44 (NSPD 44) which authorizes the creation of civilian
capabilities for stabilization and reconstruction. True enough, the words “insurgency”,
“insurgent” or “counterinsurgency” do not appear in NSPD 44, but it clearly envisages
the need to deploy integrated whole-of-government capabilities in hostile environments.
Personnel policies to develop human capital also require effort, but might be less of a
burden than we currently envisage. Rather than sweeping policy changes, we simply need
relatively minor modifications such as the ability to identify and record civilian officials
with appropriate skills for conflict environments, track them throughout their careers,
provide financial and legal cover for deployments, give them the necessary individual
and team training to operate in hostile areas, and create career structures (perhaps in the
form of “additional skills identifiers”) that recognize time in conflict zones as equivalent,
for career purposes, to time in standard postings.
Organizations, again, perhaps need less modification that we might imagine. We already
have a near-perfect instrument for inter-agency counterinsurgency in the form of the
Country Team, a 1950s innovation that has proven highly effective in adapting to
complex environments. It remains the only standing inter-agency organization in the
USG that can deliver integrated whole-of-government effects. It is thus an extremely
valuable tool that we should be working to improve even further. Other organizational
approaches, such as the Provincial Reconstruction Team (PRT), provide a basis for
adaptation. PRTs were invented in 2003 in Afghanistan and have often been treated as a
panacea for civilian counterinsurgency. They are not. But careful analysis of why PRTs
7
succeed in some areas and do less well in others can help tailor approaches for specific
situations. In this context, the efforts of private firms like Aegis Defence Services, whose
Reconstruction Operations Centres and Regional Liaison Teams are flexible inter-agency
organizations that have worked extremely well in Iraq, are worth emulating. Similarly,
while NSPD 44 envisions a civilian reserve corps deploying field personnel and middlemanagement
into conflict environments, we could also use it to establish a smaller expert
cadre of advisors who could assist Ambassadors, Country Teams or force commanders.
Systems capabilities (electronic and otherwise) require significant work. These might
include skills registers, personnel databases, and field capabilities such as
communications, transportation and protection equipment. We could also benefit from
electronic platforms to enable sharing of information between agencies, including nongovernment
organizations. ReliefWeb is a good example of this, allowing multiple
agencies to post and share information, identify opportunities to collaborate, and
deconflict efforts. Security protocols allow information to be shared only with authorized
participants, while public information can be widely disseminated. ReliefWeb’s
Afghanistan page (http://www.reliefweb.int/rw/dbc.nsf/doc104?OpenForm&rc=3&cc=afg) covers
many components of the “three pillars” model, in the context of a complex emergency.
Building on this would be less difficult, and less expensive, than one might think.
Training and education (for civil, military, and non-government personnel) would also
create shared understanding, and spread best practices throughout a “counterinsurgency
community” — again helping us achieve collaboration across a wide variety of players
whom we cannot control. Besides specific educational outcomes, these programs develop
personal relationships and erode institutional paranoia. Specific training needs include
the development of civilian teams capable of “early entry” into environments not yet
secured by military or police forces, with the movement, communications and selfprotection
skills and equipment to operate in these areas. Other needs are a capability for
“denied area ethnography” to collect human terrain and population data for effective
planning, and education for military leaders in the significant body of expertise that aid,
humanitarian assistance and development communities have built up over time.
Finally, doctrine — a common USG handbook, common funding & legal authorities, and
common operating standards — might be useful. And so we come full circle, to the
OIDP of 1962. But it should now be clear that, without a common mental model for the
environment and the pillars of a counterinsurgency effort, and without the personnel,
organizations, systems, training and education elements of capability in place, merely
producing a doctrinal handbook is likely to be as little use in 2006 as it was in 1962.
Conclusion
These thoughts are tentative; they need a large amount of work. The “three pillars” model
is clearly incorrect — all models are, in that they are systematic oversimplifications of
reality. But this, or something like it, might be a basis for further development.
And time is of the essence: regardless of the outcome of current campaigns, our enemies
will keep applying these methods until we show we can defeat them. Thus, this is one of
the most important efforts that our generation of national security professionals is likely
to attempt. Our friends and colleagues’ lives, the security of our nation and its allies, and
our long-term prospect of victory in the War on Terrorism may, in part, depend on it.
*********
8
NOTES
1
United States Department of Defense, Quadrennial Defense Review Report 2006, U.S. Government
Printing Office, Washington D.C. 2006, Chapter 1 for a detailed exposition of this argument.
2
See State Department, Office of the Historian, Foreign Relations of the United States, 1961-63, Volume
VIII, Document 106, U.S. Government Printing Office, Washington D.C., 1990, pp. 382-383
3
Ibid. p.3
4
Ibid, Document 105.
5
Charles Maechling, “Camelot, Robert Kennedy, and Counter-Insurgency: A Memoir”, in The Virginia
Quarterly Review, at http://www.vqronline.org/printmedia.php/prmMediaID/7976 accessed Sep 06.
6
Ibid.
7
See Robert R. Tomes, “Relearning Counterinsurgency Warfare” in Parameters, Spring 2004, pp.16-28.
See also John A. Nagl, Learning to East Soup with a Knife: Counterinsurgency Lessons from Malaya and
Vietnam, 2nd
Edition, University of Chicago Press, Chicago, 2005; Robert M. Cassidy, “Back to the Street
Without Joy: Counterinsurgency Lessons from Vietnam and Other Small Wars” in Parameters, Summer
2004 pp. 73-83; and N. Aylwin-Foster, “Changing the Army for Counterinsurgency Operations” in Military
Review, November-December 2005, pp. 2-15.
8
Including, but not limited to, the Department of State, Agency for International Development, Department
of Homeland Security, Department of Justice, Federal Bureau of Investigation, Department of Treasury,
Central Intelligence Agency, Department of Commerce (International Trade Administration), Department
of Defense, Department of Energy (National Nuclear Security Administration), Department of Labor
(Bureau of International Labor Affairs), International Trade Commission, National Security Agency,
National Security Council, United States Trade Representative, etc etc.
9
This definition follows that put forward by Gordon H. McCormick, who suggests that “an insurgency is a
struggle for power (over a political space) between a state (or occupying power) and one or more
organized, popularly based internal challengers”. (McCormick, “Things Fall Apart: The ‘Endgame
Dynamics of Internal Wars”, RAND, draft paper, forthcoming, p. 2). But, to take into account the transnational
nature of several contemporary insurgencies, I have replaced McCormick’s notion of a single state
entity facing an internal challenger with the broader concept of a state or group of states confronting one or
more (internal or external) non-state challengers.
10
I am indebted to Dr. Gordon McCormick of the Naval Postgraduate School, Monterrey and to Colonel
Derek Harvey for insights into the “small world, scale-free” aspects of insurgent social networks and the
enduring influence of the pre-war Iraqi oligarchy on the current Iraqi insurgency.
11
Including psychological operations, electronic warfare, computer network operations, military deception
and operations security. See U.S. Department of Defense, Joint Publication 3-13 Information Operations,
13 February 2006.
12
For a description of this approach, see United States Agency for International Development, Fragile
States Strategy, January 2005, PD-ACA-999, USAID, Washington D.C. 2005, p.3 ff.
13
See United Nations, Department of Peacekeeping Operations, Disarmament, Demobilization and
Reintegration of ex-Combatants: Principles and Guidelines, United Nations, New York, 1999, available
online at http://www.un.org/Depts/dpko/lessons/DD&R.pdf
14
See The White House, “President Bush Addresses United Nations General Assembly” at
http://www.whitehouse.gov/infocus/mideast/ accessed September 2006.