Last War’s Theories in a New Conflict: Testing Civil-Military Relations Theories on the Afghanistan War (2001-2018)

(1)

Last War’s Theories in a New Conflict

Testing Civil-Military Relations Theories on the Afghanistan War

(2001-2018)

Frank van Zeggeren

Student ID: 2106329

Supervisor: Dr. M. Popovic

Second Reader: Dr. T. Abbas

Master Thesis, 10-01-2021

Crisis and Security Management

Wordcount: 22,715

(2)

(3)

1. Introduction

1.1. Civil-Military Relations Theory and Afghanistan

In modern liberal democracies like the United States of America there appears to be an expectation that the military, despite all its violent power, will always follow the instructions of its civilian political leaders. Indeed, an extreme situation in which the military completely ignores the democratically elected government or even overthrows it in a coup is very unlikely in the US. Still, in more subtle ways, the military can attempt to not fully follow its instructions given by the civilian leadership, or try to influence said civilians in order to manipulate the instructions in accordance to its own preferences. The conflict that presents itself here is the possible clash between the preferences of civilian leaders and the military that has to perform as instructed. This interrelation is often called the Civil-Military Relation (CMR) and has spawned its own field of study by scholars from various backgrounds. This academic field attempts to make sense of the way in which a society and its military deal with each other in a very broad range of topics.

Scholars in this field have attempted to create theories to explain the ways in which this relationship develops. One of the topics in which such theories have been formulated is the issue of conflicting preferences in military operations. Two of the more prominent theories in the field that are relevant to this topic have been proposed by Deborah Avant and Peter Feaver respectively. In their theories, the authors both base their analysis on the principal-agent framework, which frames the civilian leadership as a principal that authorizes the military to act as its agent. However, the exact application of this theory differs significantly between the two theories.

Avant applies the principal-agent framework to the institutional structure of the US and its implications for military actions, arguing that there are in fact two principals whose division drives military actions. Feaver, on the other hand, uses the theory to analyse the strategic nature of the interaction between the civilians and the military, arguing that each actor’s actions are often based on calculations on what their counterpart will do. As will be examined in detail in the Body of Knowledge section of this thesis, this difference in application has significant consequences for their explanations of behaviour in the civil-military relation. As such, both authors have also criticized each other’s approach as being mostly suitable for analysis but missing a fundamental component which is better covered under their own respective theory.

(6)

2

One other element that these two theories have in common is that they both have been formulated based on analysis of events in the 1990’s. After an accusation that the military had achieved too much dominance over the civilians in this decade, both Avant and Feaver set out to analyse the relationship more thoroughly, leading to their theories. In doing so, they based their analysis on conflicts in the 1990’s, such as Haiti, Somalia, and Bosnia. While these cases were relevant for the discussion at the time, they do not necessarily provide a good example for other military conflicts. For example, in the 1990’s the US engaged in these conflicts voluntarily as they wanted to contribute to humanitarian missions which would only last for a limited amount of time. This is in contrast to a war which is started out of necessity, for example after being attacked, or a war that lasts for multiple years or even decades. It is therefore possible that the theories that explain the civil-military relation in the wars of the 1990’s do not explain the relation in other conflicts.

It is important, therefore, to test these theories on a more recent conflict to validate the explanations on the civil-military relation that the authors provide. The objective of this thesis will therefore be to test Avant’s and Feaver’s theories by applying their expectations to the case of the Afghanistan war. By analysing events and processes of this war and then comparing them to the mechanisms prescribed in theories by Deborah Avant and Peter Feaver, this thesis will attempt to establish the degree of their applicability and explanatory value. If successful, this thesis can determine to what extent the described processes and corresponding expectations formulated in these theories remain relevant in a case outside of those cases used in their formulation.

Since these theories are quite broad and this analysis is limited in its scope, this thesis will focus on one element of both theories: the circumstances under which the military is able to achieve its own preferences for policy for the Afghanistan war. Based on the theoretical work from both Feaver and Avant, expectations can be formulated which postulate how certain conditions and actions within the civil-military relation will influence the degree to which the military is able to attain its preferences in official policy. These expectations will be tested in this thesis on the case of the Afghanistan war from 2001 through 2018.

The case of the Afghanistan war has been selected due to its difference in characteristics to the wars of the 1990’s. Firstly, the Afghanistan war started not out of choice but out of necessity after a terrorist attack on US soil. Secondly, the war has lasted almost twenty years, as opposed to short humanitarian missions in the 1990’s. The war also provides a case where the relationship between the civilians and the military has regained public attention. As the war has

(7)

3

continued to drag on, more people have started to analyse the interactions between the two actors and how they affected the outcome of the war. By applying the two theories to this case, this thesis may also provide additional understanding to a topic that remains highly relevant in public and political discussions.

1.2. Research Question and Structure

To achieve this thesis’ goal of testing Avant’s and Feaver’s theory by applying them to the case of Afghanistan, the following research question has been formulated:

To what extent do the theories proposed by Deborah Avant and Peter Feaver explain the degree to which the military was able to achieve its preferences for policy in the Afghanistan war (2001-2018)?

To answer this research question, the thesis will proceed as follows. First, the thesis will determine the expectations for military preference attainment as formulated by Avant and Feaver. This will be achieved by examining the literate in the Body of Knowledge section, and then from these theories expectations will be formulated that can be analysed. Next, the thesis will attempt to establish the circumstances of the Afghanistan war, specifically the circumstances under which the military did or did not attain its preferred policy options for this conflict. This will be done in the four empirical chapters, by carefully examining events that transpired and presenting them in a manner that makes them suitable for analysis. It will then be determined whether the formulated expectations correctly explain the events that have been described, by continuously analysing each empirical chapter through these theories. Finally, the combined results of this analysis will be presented in the conclusion, where the main research question will be answered.

The thesis will be consist of a longitudinal case study design. The time frame of 2001-2018 has been selected as to be able to capture as much of the war as possible, while also maintaining a sufficient number of available sources from which to draw evidence. The sub-divisions of chapters have been selected in chronological order, but not by selecting the different presidents and their administrations. Instead, the chapters are divided into periods with roughly similar policies in Afghanistan.

The thesis consists of eight chapters, this introduction being the first one. The second chapter will address the available relevant literature, and will state how the relevant theories are formulated. The third chapter consists of the methodological explanation of the analysis of this thesis. Chapters four through seven then analyse the Afghanistan case in four different time

(8)

4

periods. Each chapter will set out events as described by the sources, as well as provide an analysis of how the described theories apply in that specific time period. Chapter eight consists of the summary of findings and the conclusion. Lastly, the sources that have been used for this research can be found at the end of this thesis.

(9)

5

2. Body of Knowledge

The purpose of this thesis is to test the applicability of theories of Civil-Military Relations by applying them to the case of the Afghanistan war (2001-2018). For this research, two theories, proposed by Deborah Avant and Peter Feaver, have been selected that are both based on principal-agent theory. Since the two selected theories are quite broadly formulated, they will be tested on specific aspects of their contents, as will be further explained below.

2.1. Literature Review: Civil-Military Relations and Agency Theory

Civil-Military Relations (CMR) is a field of study that focusses on the question of how any society manages the military, which is needed to protect society from foreign threats, without this branch turning against the society it is supposed to protect. In its most blunt form, this provides the dilemma of giving the military enough power to protect society, without the military turning against its civilian leaders in some sort of coup. The overpowering of society by the military can also manifest itself more subtly, in that a military that is strong enough to repel threats may cost so much to the society that it can no longer afford its many other civilian functions.1_{In other words, ‘the very institution created to protect the polity is given sufficient}

power to become a threat to the polity.’2_{Adding to this is the fact that in order to succeed in}

warfare, the military needs to have a very distinct set of values and organisational structures. These values and structures will be at odds with their civilian counterparts, but must never be allowed to negatively influence the latter.3 The management of these apparently conflicting interests and values within a society is what lays at the heart of civil-military relations.4

In western democracies, the question of a coup appears less relevant. Instead, authors who study the topic focus more on the question to what extent and under which conditions the civilian leadership has control over its military in all its actions, ranging from actual warfighting to subjects such as materiel acquisition or who gets to join the military. In these cases, the literature often examines the civilian control of military actions, where there is general consensus on the normative principle that the civilian should be in control of the military, and not the other way around.5

1_{Hiroshi Nakazato, “Defensive Discourses: a Narrative Theory of Civil-Military Relations” (dissertation, 2006),}

4-5.

2_{Peter Feaver, Armed Servants: Agency, Oversight, and Civil-Military Relations (Cambridge, MA: Harvard}

University Press, 2005), 288-290.

3_{Nakazato, Defensive discourses, 4-5.} 4_{Feaver, Armed Servants, 4.}

(10)

6

The field has produced a broad variety of works, using different analytical approaches. One of the first major contributors to the field was Samuel P. Huntington, whose book The Soldier and

the State provided the foundation for many of the discussions and contributions that would

follow. Huntington wrote his book in 1975 in response to the Cold War, and in an attempt to provide the United States with the framework to win it.

To summarize the main arguments very shortly, Huntington argued that there should be a very strong distinction in the civil-military equation. The military is claimed to be fundamentally distinct from the civilian in terms of his professionalism and military ethic, which makes it suited for his military tasks. The civilian leadership is then argued to be characterized by the type of control it uses and its political ideology. This civilian control of the military, Huntington argued, can take two forms: subjective and objective control. Subjective control would entail that the civilian would attempt to have the most power over the military in society. Objective control would entail that the civilian would push for the most amount of professionalism in the military by staying out of military affairs, with the trade-off that the military would agree to being a tool of the state. Huntington would argue for the latter option of control.6

After Huntington, many schools of thought followed within the field of CMR, but all responded to Huntington’s work and his framework of subjective and objective control. After a perceived ‘crisis’ in the relationship between the military and the civilians in the 1990’s, the field regained some of its prominence and produced a new wave of publications. These publications reimagined the relationship by turning not to the concepts that are to be examined but rather the processes by with the relation was formed.7

To investigate these relations, authors turned to principal-agent theory. As used by CMR authors, principal-agent theory states that within any organisation, from a company to society as a whole, the actor in charge, or principal, will delegate certain tasks to an agent who works in his name. This creates a problem, however, since the agent may do something that the principal never intended – this is referred to as a problem of agency. This may be because the interests of the principal and agent are not the same. It may also be the case that, since the agent has specialist knowledge of the delegated task and a better view of what is actually happening, a problem of hidden action or hidden information occurs. This means that the agent can turn certain situations in its favour by not telling the principal all relevant elements of the task. The

6_{Nakazato, Defensive discourses, 32-33.} 7_{Idem, 43-44}

(11)

7

principal will therefore have to figure out ways in which it can monitor the agent, to make sure it performs as intended.8

One of the authors who applied this framework to CMR is Deborah Avant. She first positions the president as the principal and the military as the agent. She then further examines the institutional make-up of the American political system and argues that there are in fact two principals in this framework: the president and Congress. These two actors do not always agree, as they are in fact designed to be in disagreement. This, Avant argues, creates problems of a divided principal, which is expected to lead to certain behaviours of the agent, in this case the military. If there are two conflicting preferences within the divided principles, the military agent is more likely to side with the one that matches its own preferences. Furthermore, even if Congress and president agree on policy, they may disagree on methods of monitoring the agent. This will often lead to giving less power to the agent, who can then only make conservative steps. This all leads to Avant’s argument that the military is more likely to give conservative estimates for military options when the civilian principal is divided, and that the military’s influence on policy also increases when the principal is divided. Conversely, if the principals are united, it is expected that the military will be more inclined to listen to civilian instruction and not give conservative estimates.9

A second author who used agency theory on the questions surrounding CMR is Peter Feaver. Feaver focusses less on the institutional aspect of the relation like Avant, but rather analyses the strategic nature of interaction between the civilian principal and the military agent. It is strategic, Feaver argues, since both the military and the civilians do not just base their actions on their own preferences, but also on what they expect their counterpart to do. Feaver’s argument in its most basic form states that both the civilian and the military have a choice between two options when a specific policy has been set out. The civilian can choose to monitor intrusively, meaning a large amount of interference and checks on the military to make sure it does what the civilian wants, or non-intrusively, which leaves more room for the agent but also increases the risk of hidden action and hidden information. The military can choose to do what the civilian principal wants it to do, what Feaver calls working, or it can choose to perform its tasks how they themselves think it should be done, which Feaver calls shirking. Feaver urges

8_{Deborah D. Avant, “Are the Reluctant Warriors out of Control? Why the US Military Is Averse to Responding}

to Post-Cold War Low-Level Threats,” Security Studies 6, no. 2 (1996): pp. 51-90, 54-56.

(12)

8

that shirking does not mean that the military does nothing or avoids responsibilities, but rather performs actions according their own preferences instead of those of the civilian.10

Feaver captures these choices and the way they interact with each other in a model which provides certain expectations. Feaver argues that the choices of intrusive/non-intrusive monitoring and working/shirking are made based on other considerations: for the civilian, the cost of monitoring and the expectation of the military shirking, and for the military the question whether civilian plans converge with their own interests and their expectations of being punished for shirking. These factors interact. So, for example, when the cost of monitoring goes down, the civilian will monitor more, thereby increasing the chance of punishment if the military shirks, therefore leading to the expectation that the military will shirk less.11

Both Avant’s and Feaver’s theoretical models stem from a particular set of occurrences in the 1990’s, dubbed a ‘crisis’ in civil-military relations at the time. Shortly summarized the main argument by those who called the situation a crisis was that the military had gained too much influence over the civilian leaders.12_{Both Avant and Feaver set out to analyse this claim and}

its nature. By focussing on different aspects of agency theory and the civil-military relation, they have some large overlap but also several distinctions. Avant, by focussing on institutional relations, places more emphasis on regular advice from the military and the make-up of American agencies. Feaver, with his attention to the strategic nature of interaction, examines the gritty realities of the relationship behind closed doors.

Both approaches have merit but also drawbacks, as they themselves have argued about each other’s approach. Feaver, while agreeing with parts of Avant’s analysis, criticizes the focus on the division between Congress and the president. He states that this division is still limited, since there are many more divisions within and between civilian and military actors. If we do have to make a distinction to apply principal-agent theory, he argues, it is best to divide between civilian and military all together.13 Avant, while also acknowledging the contributions by Feaver’s approach, takes issue with his focus on friction. By putting so much emphasis on shirking and punishment, it does not actually prove who is in control. If there is no friction,

10_{Feaver, Armed Servants, 2-4.} 11_{Idem., 112.}

12_{Deborah Avant, “Conflicting Indicators of ‘Crisis’ in American Civil-Military Relations,” Armed Forces &}

Society 24, no. 3 (1998): pp. 375-387, 375-378.

(13)

9

does that mean there is complete control by the civilian over the military or has the military completely dominated the civilian? A focus on friction alone cannot provide an answer.14 The two theories are, however, both based on specific events in the 1990’s. In these conflicts the US was not attacked, but rather chose to engage in conflicts around the world. These conflicts were relatively short-term and could perhaps be seen as a lesser variant of large-scale warfare. The question then arises whether these theories can also be applied to a case in which the US was attacked, and the resulting war was both involuntary and long-term.

2.2. Analytical Framework: Applying the Theories

Therefore the question becomes whether the two theories that were designed in a time of (relative) peace for the US can still be applied to a case of long-term war, provoked by an attack on the US itself. To test this applicability, this thesis will take some main elements of both theories in regard to the military’s ability to influence policy and try to test them to the case of the Afghanistan war, in the period 2001-2018. Since both theories are quite broad and this thesis limited in scope and size, two main expectations are formulated for the two theories, which will then be tested in the empirical section.

Avant argues that a divided principal will lead to several expectations. Two of these expectations will be combined into one: the expectation that a divided principal will lead to conservative advice and action by the military; and the expectation that a divided principal will increase the influence the military has on policy making. A divided principal is portrayed by Avant as a situation in which Congress – opposing or ruling party’s members or leader, or head of a relevant Congress (Sub-)Committee on Foreign Affairs or Armed Forces – publicly voices an opposing opinion or attitude to that of the executive: the president and his administration.15

The expectation put forth by Avant is that such a division will create different, possibly contradicting demands made by principals, which the military agent will have to adhere to. In order to do so, the military is expected to give ‘conservative’ advice and action, meaning it will take slow, well supported steps, as to not get blamed by either principal if things go wrong.16 Combining these two elements from Avant’s theory, the first expectation which will be tested in this thesis is:

14_{Avant, “Conflicting Indicators,” 382.} 15_{Avant, “Out of Control?” 63.} 16_{Idem., 58-59.}

(14)

10

The military will be better able to turn its conservative preferences into policy when the civilian principals are divided; if the civilian principals are united, the military will be less able to do so.

Feaver’s theory provides a very extensive range of different expectations, based on civilian and military behaviour. Since this thesis focusses on the perspective of the military and its ability to influence policy, a choice has been made to formulate an expectation based on relevant elements in this regard. To apply Feaver’s theory in this thesis, one step has to be explained. Feaver focusses on actions by the military, whether it works or shirks, and not explicitly whether the military advises or influences policy. He does, however, incorporate these elements in his framework, although it becomes somewhat hard to pin down for analysis. His point is that shirking can take on many forms, and definitely does not just entail simply disobeying orders. By leaking to the press, giving inflated estimates of cost for certain actions, or being unrealistically slow in the policy process, the military can attempt to change the policy the civilian wants.17

The effectiveness of these attempts to change policy is not explicitly mentioned in Feaver’s framework, but can be considered a product of bargaining over policy. This ‘bargaining’ needs some more clarification in order to place it into this thesis’ framework. If the preferences for policy by civilians and the military do not align, the civilians will have to formulate policy that the military can execute, but also that they themselves will politically survive. If the civilian simply pushes for its own preferences, the military can shirk by leaking to the press or other means to inform the public and political opponents of the principal. In theory this does not have to matter, but in practice this can have enormous political consequences for the civilian principal. For example, if the civilian principal is politically ‘weak’, punishing a shirking military may cause a significant political backlash. Fearing such a political cost, the civilians will have to persuade the military to publicly support its proposed policy, by bargaining and incorporating some of the military’s preferences.18

The relevant question then becomes whether the military decides to shirk, whether the principal punishes said shirking, and whether the military can achieve its preferences through shirking. Feaver’s model suggests that the military makes its decision on whether to shirk based on two factors: how much difference there is between the civilians’ preference and their own; and the

17_{Feaver, Armed Servants, 68} 18_{Idem, 96-101.}

(15)

11

expectation the military will to be punished for shirking.19 Secondly, since a politically weak principal is less likely to punish its military agent for shirking, it is expected that the civilian will have to make more compromises in its policy to acquiesce the military agent. Feaver’s expectations are therefore that the military will shirk more when it expects to not be punished, and that the longer the shirking goes unpunished the more influence it can attain on policy. Therefore, the expectation which will be tested in this thesis is:

The military will attain more influence on policy through bargaining, when military shirking goes unpunished by the civilian principal; if shirking is in fact (expected to be) punished, the military will attain less influence.

So, to conclude the first step of this thesis, the expectations formulated by Avant and Feaver regarding military policy preference attainment can be set out in the following table:

Theory Expectation Elements to analyse

Avant The military will be better able to turn its conservative preferences into policy when the civilian principals are divided; if the civilian principals are united, the military will be less able to do so.

- Is there a divided principal? - Is the military preference

conservative?

- Is the military’s preference made into policy?

Feaver The military will attain more influence on policy through bargaining, when military shirking goes unpunished by the civilian principal; if shirking is in fact (expected to be) punished, the military will attain less influence.

- Is the military shirking? - Is it unpunished?

- Is the military’s preference made into policy?

(16)

12

3. Methodology

3.1. Research Design

The goal of this thesis is to test the validity of theories of civil-military relations that were made based on events in the 1990’s, by using them for analysis on a case which occurred after the theories’ formulation. As stated above, the two theories by Avant and Feaver have been selected due to their relevance to the field of civil-military relations, their comparability in regard to principal-agent theoretical underpinnings, but also their distinct differences in term of focus of analysis and units to examine. To test the two theories a deductive research will be performed in this thesis. To answer the main research question the research will use a qualitative longitudinal case study design. This type of design has been chosen since it enables a deeper investigation of variations of the studied actors and their interactions over time. By structuring the research in this manner, the thesis will be able to better address the ways in which change over time is achieved and influenced within a case.20_{In this thesis, that single case is the}

Afghanistan war, and the different time periods have been selected based on the shifts in official policy regarding that war.

Selecting this design has an advantage over multiple case studies, since the latter would require a complete evaluation of local contexts of each conflict, as well as the different political calculous made for each of those conflicts by the civilians and military. By using the same conflict, more elements of the conflict remain constant, enabling a better focus on the changes within the civilian and military leaderships’ assessments and interrelations.

The data for this thesis will be collected by gathering several types of primary and secondary sources related to the war in Afghanistan. Primary sources consist of speeches, congressional hearings, official publications and declassified documents. Secondary sources will consist of historical works, memoirs, journalistic publications, and other sources provided by authors and analysts of these events. Analysis of the sources will be done according to the historic research method. Since the theoretical underpinnings of this project are heavily reliant on interpretations of events that are not openly discussed due to their secretive nature, it is important that a good contextual analysis is provided to make sure events are not misconstrued. Qualitative historical

20_{Bren Neale, “Researching Educational Processes through Time: the Value of Qualitative Longitudinal}

Methods,” in Handbook of Qualitative Research in Education, ed. Michael R. M. Ward and Sara Delamont (Cheltenham: Edward Elgar Publishing, 2020), pp. 102-114, 102-103.

(17)

13

analysis can provide a detailed account of events in order to better categorize these cases for research and testing of theories.21

3.2. Case Selection and Timeframe

To achieve the thesis’ goal of testing theories on civil-military relations a case must be selected that is a conflict in the twenty-first century, and also a conflict in which the motivation to go to war was fundamentally different than in the 1990’s – i.e., retaliation for an attack on the US instead of voluntary participation in a humanitarian mission. The Afghanistan war is selected as the case to be studied since it meets these criteria. The Afghanistan war can be argued to be a typical case, since it falls within the scope of both theories. Regarding Avant’s theory, the war presents clear episodes in which Congress and the executive disagree over policy. In the case of Feaver, the Afghanistan war fits within his general framework for civil-military relations in times of war. The case is also well suited for a longitudinal study due to its long duration, thereby enabling all the analytical advantages described in the research design section.

The timeframe for this thesis has been established from 2001 through 2018. The start of the timeframe follows somewhat naturally from historical events. The year 2001 is selected as a starting point as it covers the preparation for the Afghanistan war, which started in that year. The ending of the timeframe is somewhat more arbitrary, since the war itself has continued after this date. The endpoint has been selected to be as late as possible, so as to capture as much as possible the variation of civil-military relation over time. By selecting 2018, the thesis attempts to achieve this aim by covering the first two years of the Trump administration. However, the timeframe in which reasonable analysis can be achieved is constrained by the availability of sources that provide insight into the studied topic. It has been decided for this thesis that after 2018 there is, at the moment of writing, an insufficient number of sources to reach meaningful conclusions on the relationship between the civilian and military leaderships. The timeframe has then been subdivided into four chronological chapters, in which further subdivision are made. The chapters are divided into time periods which have roughly similar military policies. A choice has been made to not divide the timeframe into political administrations, so as to highlight the progress of policy throughout different administrations, and the changes that occurred within these administrations. This has been done to step away from an analytical focus on political affiliation alone, and rather focus on the effects on actual

21_{Cameron G. Thies, “A Pragmatic Guide to Qualitative Historical Analysis in the Study of International}

(18)

14

policy. The sections within the chapters are somewhat chronological, but mostly focus on relevant policy elements within the chapter. As such, sections can somewhat overlap in timeframe but analyse a different element within that period.

3.3. Operationalization

In order to analyse the different elements of the proposed theories and how they may or may not apply in the case studied in this thesis, it is important to formulate what exactly constitutes these elements within a real-world scenario.

The first concept to be explained applies to both authors, as they both attempt to determine under what conditions and to what extent the military is able to attain its preferences for military policy, in this case in Afghanistan. First, then, it needs to be established what constitutes military policy in this case, how to determine what military preferences are, and how to determine whether these preferences were attained in officially formulated policy.

Military policy

Military policy in this thesis will be understood as policy sanctioned by the civilian political leaders in regard to military affairs in Afghanistan. In the context of the Afghanistan war, this will mostly concern the strategy or way-of-war approach used in this conflict, e.g., counterinsurgency, counterterrorism, invasion and occupation of a foreign country, etc. It will also relate to the overall goal of the conflict, for example, taking out Al Qaeda only or also the Taliban, getting out as fast as possible or setting clear goals for development of Afghanistan before a withdrawal. It will also concern itself with the number of troops deployed to the country, and the duration for which there will be resources to fund and equip these forces. All in all, it encompasses a wide variety of strategic and operational decision that are ultimately decided by the civilian and publicly presented as policy.

Military preferences for military policies

While the civilian political leadership has its own ideas on warfare and ultimately decides, the military of course has its own preferences. Preferences are somewhat more difficult to determine, as they are not always made public for a variety of reasons. For this thesis, military preferences will be understood as those military policies that are actively bargained for by military leadership. This latter term, military leadership, creates its own problems as the military leaders are sometimes themselves divided. For clarity, this thesis will refer to the Chairman of the Joint Chiefs of Staff (CJCS) when talking about military leadership. This position functions

(19)

15

as the primary military advisor to the president and the Secretary of Defense, and therefore it makes a logical option for being used as an indicator for the military leadership. If any other military actor is meant in this thesis, this will be mentioned accordingly.

Attaining preferences in policy

Building upon the operationalization of military preferences for military policies, the attainment of such preferences is understood in this thesis as those situations in which the defined military preferences can be clearly recognized in official policy, or those policies on which it was commented by involved actors that the military achieved its goals.

For the theories, several key elements need to be explained in order to be recognized within the case.

Divided principal

In Avant’s theory, the principal-agent framework in civil-military relations actually has two principals: Congress and the executive, i.e., the president and his cabinet. These principals are united when Congress mostly supports the president and his policies, or at least seems to vote in favour of his policy. The principals are considered to be divided when congressional leaders or heads of relevant congressional committees, such as the Armed Services Committees, take a public stance against policy by the executive.22 Criticism can of course be voiced rather quickly, so it is important to determine whether this criticism has broader support in Congress beyond just one member of Congress voicing it.

Conservative assessment or action by military

A next step in Avant’s theory is the conservative assessment or action. In the theory, when principals are divided they will make different and possibly contradictory demands of the military. The military will want to avoid blame by any principal after the fact. Therefore, its advice and action will consist of small steps that signal their attention to concerns of both principals. This process will ensure that their bases are covered by official approval, and thereby covering in for blame if the policy does not result in the required outcomes.23

This is a rather broad concept, and it is highly dependent on the context of a situation in which the military operates. In the case of this study, this can take the form of advice to not suddenly

22_{Avant, “Reluctant Warriors,” 62-63.} 23_{Idem, 58-59.}

(20)

16

increase or decrease troops levels, not suddenly but incrementally change strategies, and other policy advice and action that appears to be somewhat of a compromise. It should be noted that conservative here does not necessarily mean to change nothing at all in a certain context. If one principal wants to change everything, and the other nothing, the military would choose a side by stating they want to change nothing. A more conservative position would then be to suggest incremental change.

For Feaver’s theory, some other concepts need to be expanded upon to be used in analysis.

Working

Feaver goes into great detail to explain what working and shirking means in his framework. Put very shortly, according to Feaver, working is doing things the way civilians want. In the context of the civil-military relation in the Afghanistan case, this will mean that the military performs the instructions given by the civilians to the best of its abilities. It can give advice on how they think any task should be done, but once the civilian has decided what it wants, the military should perform its tasks as instructed by the civilian. An important additional element in the civil-military context is that working means not only performing the task according to the civilians’ satisfaction, but doing so in a manner that protects the civilian from external threats, while also making sure that the civilian remains in control.24

Therefore, in the case of the Afghanistan war it not only means to fully implement the instructions given by the civilian, but also ensuring that the military does not in any way try to sneak in its own preferences without explicitly telling the principal, as it would do in its advisory role. Methods of this kind of unwanted influence are better described in the opposing concept: shirking.

Shirking

Shirking is when the military does not follow the instructions of the civilian by not doing what the civilian wanted, not in the way civilians ordered it, or in such a manner that it frustrates civilians in their efforts to make other decisions. In the context of civil-military relations in the US, specifically Afghanistan, this will not take on the form of refusal of orders. Rather, the military can inflate the cost in lives and money it projects an operation would cost, leaking to

(21)

17

the press to stir up political opposition to the civilians’ plans, or delaying a policy to make sure it is never actually put into practice.25

Punishment

When a civilian principal decides that a military agent has not worked but shirked, it can enact certain policies or actions that ‘punish’ the military. This can be done to increase the likelihood that the military agent will work in the future instead of shirking again. Examples of punishment can be more intrusive monitoring, meaning the military will have less room for initiative; cutting budgets; firing or otherwise removing a military member from the service; military justice; and what Feaver calls extralegal civilian action, which encompasses everything from public oral rebukes to Stalinist purges of the military.26

Expectation of punishment

This is perhaps the most difficult concept to pin down in Feaver’s framework. In essence, there is a wide variety of factors that lead the military to expect punishment by the civilian. Like the conservative estimate and action in Avant’s theory, it is highly dependent on the context of the situation. Feaver does name some general reasons for a certain expectation of punishment, as he states that a popular president is more likely to fire an unpopular general, than an unpopular president is likely to fire a popular general. Likewise, Feaver remarks that unified principals are more likely to punish the military than divided ones, and united military agents are less likely to be punished than divided ones.27

Within the case, therefore, the degree of expectation of punishment has to be argued through contextual events. For example, there can be signs that a principal is willing to punish, by punishing the military in the past. Or it needs to be argued that a civilian principal is less likely to punish a military agent, because the principals are divided on policy.

25_{Feaver, Armed Servants, 68.} 26_{Idem, 90-94.}

(22)

18

4. 2001-2006

Having examined the theories and how this thesis will attempt to test them, we can now turn to the case at hand. Throughout the next chapters, events surrounding the Afghanistan war will be first portrayed in some detail and then examined along the theories to be tested. At the end of each chapter a preliminary analysis is made of the events in the selected time period. This analytical process of the case begins in 2001.

4.1. 2001-2002: Invasion and Rumsfeld’s Transformation

The terrorist attacks on 11 September 2001 would form the starting point of the United States’ war in Afghanistan. After several airplanes were hijacked and flown into the World Trade Centre, the Pentagon, and a rural field, the US quickly set out to militarily respond to those responsible. However, the relationship between the military and the civilian leadership had at this point already been heavily strained, since the Secretary of Defense Donald Rumsfeld was seen as extremely overbearing to his generals. Rumsfeld thought that the military had to be brought back ‘under control’ after the Clinton administration, and also pushed for military reforms. Rumsfeld envisioned a transformation of the military, in which new technologies would enable it to get the same results in less time and with fewer resources. The military itself was not convinced of either point, souring the relationship.28 However, after much confrontation, the military was seen by some to come out on top as Rumsfeld toned down his reform goals.29

After the attacks were carried out, the national security establishment quickly came together to discuss the situation and their options to retaliate. Once it was established that Al Qaeda, a terrorist group mostly based in Afghanistan, was responsible, discussions began on how to proceed. It quickly appeared that the military did not have any plans readily available for Afghanistan. The military had a plan to only use cruise missile strikes, or to being a months-long preparation to deploy multiple divisions of troops. These plans were seen as too weak or too slow by the civilians.30

28_{Russell F. Weigley, “The American Military and the Principle of Civilian Control from McClellan to Powell,”}

The Journal of Military History 57, no. 5 (1993): pp. 27-58, 27-30; Lyle J. Goldstein, “General John

Shalikashvili and the Civil-Military Relations of Peacekeeping,” Armed Forces and Society 26, no. 3 (2000): pp. 387-411, 390-392 Richard F. Kohn, “Out of Control: The Crisis in Civil-Military Relations,” The National

Interest, 1994, pp. 3-17.; and Aaron Major, “Which Revolution in Military Affairs?,” Armed Forces & Society

35, no. 2 (2009): pp. 333-361, 333-334.

29_{Feaver, Armed Servants, 290.}

30_{James H. Lebovic, Planning to Fail: The US Wars in Vietnam, Iraq, and Afghanistan (New York: Oxford}

University Press, 2019), 129 and Bob Woodward, Bush at War: Inside the Bush White House (New York: Simon & Schuster, 2002), 25.

(23)

19

The civilian leadership wanted a far more aggressive response. As the military did not appear to have options that satisfied the civilians, their attention soon turned to the CIA. The intelligence agency had been active in the country since 1998, after an earlier terrorist attack on US embassies.31 The CIA suggested a plan of retaliation through paramilitary forces that would link up with local rebels, and would be supported by military Special Forces and Air Force bombings. The plan was seen by some as unrealistic, but it was exactly what President Bush wanted and was consequently approved. The CIA thereby became a major player in war planning, which had so far been a military affair.32

The military response to this shift was somewhat ambivalent. The military seemed bitter about their lost dominance of war planning, and its leaders seemed confused at times on who was actually in charge of the operation.33 However, they did not really seem to object to the proposed plan. The idea of a ‘light footprint’, or few troops on the ground, assuaged their fears of getting bogged down in Afghanistan like the Soviets had before them.34_{Nevertheless, the relationship}

between the uniformed military and its civilian head remained sour. General Franks, commander of the Afghanistan campaign, allegedly found Rumsfeld so overbearing that he threatened to resign.35

Another driver for the decision to use the CIA plan was that it could start rapidly, in contrast to the military plan that required a long build-up of forces. Bush felt pressure from an American public that wanted revenge, but was also informed that public support for military action would be strongest when it was started quickly.36 He had stated that US public would only be patient for so long before there needed to be visible retaliation.37 Indeed, public opinion polls showed overwhelming support for military action in October 2001.38 Accordingly, but more significantly for this thesis, Congress approved the military response against those responsible almost unanimously.39

31_{Walter L. Perry, and David Kassing, Toppling the Taliban. Air-Ground Operations in Afghanistan, October}

2001–June 2002 (Santa Monica: RAND Corporation 2016) 18-22.

32_{Idem, 25-26; Woodward, Bush at War, 50-53 and 101; and Lebovic, Planning to Fail, 129.} 33_{Woodward, Bush at War, 243-247.}

34_{Idem, 193.}

35_{Michael R. Gordon and Bernard E. Trainor, Cobra II: the inside Story of the Invasion and Occupation of Iraq}

(London: Atlantic, 2007), 27.

36_{Woodward, Bush at War, 32.} 37_{Idem, 150.}

38_{Frank Newport, “Public Opinion of the War in Afghanistan,” Gallup (Gallup, October 31, 2001),}

https://news.gallup.com/poll/9994/public-opinion-war-afghanistan.aspx.

39_{Mark Pearcy, “‘Sixty Words’: Teaching About the Authorization for Use of Military Force (AUMF),” The Social}

(24)

20

As it turned out, fears of recklessness of the CIA plan appeared unwarranted as many Taliban fighters quickly surrendered. The ousting from power of the Taliban and Al Qaeda occurred at a rate that surprised even CIA and military planners. The operation was dubbed a success, with the notable exception of the escape of high-level enemy leaders, including Bin Laden.40

Looking at the events of this period through the lens of the two theories, several observations can be made. In regard to Avant’s expectations, it can be argued that the principals were united since Congress approved of the president’s conduct. This meant that the military’s only option for a conservative plan was to follow wishes from both principals. Their first plan was regarded as too slow and unaggressive, and was therefore not an attempt to appease both principals. Consequently, this military’s plan was rejected and it therefore did not attain its preferences. As such, the events follow Avant’s expectation that the military will not attain its preferences when the principals are united.

Feaver’s expectation, however, does not explain the events of this period. While Rumsfeld was considered overbearing, the military had proven resilient to his style in the past by minimizing his reforms. Expectation of punishment based on Rumsfeld’s behaviour alone does not fully explain events, since they had confronted this before and won. Still, when the military preference was rejected, the military did not shirk and instead worked. In this case, therefore, the military does not attain its preferences, but still works despite a lack of punishment or fear thereof. Of course there are other reasons why the military could have decided to follow the civilian’s plan, but these are not covered under this expectation. It could perhaps be argued that the new role of the CIA and loss of prestige and responsibility of the military in general is to be regarded as punishment. There is, however, no evidence of such a calculation by the military. Adding to this, such a form of punishment is not covered in Feaver’s theory. As such, the expectation is regarded to be not applicable to this period.

4.2. 2003-2006: Disinterest, Experimentation, and Reaffirmation

The next period to be checked for application of the theories is that of 2003-2005, when the US seemed more focussed on Iraq than Afghanistan. In early 2002, after most of Afghanistan had fallen from Taliban control, civilian leaders were quick to declare victory. Subsequently, they increasingly turned their attention to a new war: Iraq. First discussions on a possible military action Iraq had already been raised the day of the terrorist attacks, but were then put on hold for the Afghanistan invasion. Already in November 2001, however, planning for Iraq had

(25)

21

secretly begun.41 When the fighting in Afghanistan decreased significantly in 2002, preparations for Iraq took ever greater forms. As such, resources were increasingly allocated to Iraq instead of Afghanistan.42

The reduction in troops and resources came with a fitting strategy. Instead of conducting large scale military operations, the US force in Afghanistan would perform certain counterterror (CT) operations but would otherwise only train the Afghani security forces. The idea was in line with Rumsfeld’s military transformation, while also aligning with the broader goal of the Bush administration to not get involved in nation-building missions. As such, the military commander in Afghanistan was given the instruction to hunt and destroy remnants of the Taliban and Al Qaeda, train the Afghan National Army (ANA) to provide their own security, and work towards withdrawal as soon as possible.43

Almost immediately, the realities of the limits of this approach became visible. The training of Afghans by international partners, who Rumsfeld thought should take up this task, was lagging. Consequently, Rumsfeld ordered the Americans to train more Afghans. But the lack of Afghan security forces also meant that US forces would take a greater role in security missions, thus increasing their workload. While extra troops were provided, this was not done sufficiently for the new tasks.44

The mission in Afghanistan, however, did not change. Nor did the limited amount of civilian leaders’ attention to the conflict. Rumsfeld appeared to take an outcome-oriented approach to the operation: he had stated what he wanted – a quick withdrawal – and left the execution to the local commander. This changed significantly in 2004, when the Pentagon read in the newspaper that the new Afghanistan commander Gen. Barno had apparently changed the strategy in the country.45

Barno became local commander in 2003, and immediately saw his troops reduced from then on. He went on to establish a combined headquarters with his civilian counterpart, the US ambassador. After review of the situation on the ground, reading some older books on counterinsurgency (COIN) operations, and multiple meetings with British officers who had

41_{Bob Woodward, Plan of Attack (London: Simon & Schuster, 2004), 1-5.}

42_{Brian Neumann, Lisa Mundey, and Jon Mikolashek, Operation Enduring Freedom: March 2002-April 2005}

(Washington, D.C.: U.S. Army Center of Military History, 2005), 11 and 28-29.

43_{Neumann et al., Operation Enduring Freedom, 13.} 44_{Lebovic, Planning to Fail, 178.}

45_{David P. Auerswald and Stephen M. Saideman, NATO in Afghanistan: Fighting Together, Fighting Alone}

(26)

22

practiced COIN operations in other conflicts, Barno had set out to formulate his own strategy in the country. He published his ‘Five Pillars’ COIN approach to his troops, and publicly declared his changed strategy. Amongst other things, it ordered US troops to get out of their large bases and instead go and stay deeper in the country among the population in an effort to create more stability. At the time, with support for Taliban decreasing and support for the government apparently increasing, this approach seemed somewhat successful.46

In Washington, however, Rumsfeld did not share Barno’s enthusiasm about the COIN approach. In fact, this approach contradicted Rumsfeld’s goal of withdrawing as many troops as possible, as Barno’s COIN approach needed much more engagement of US forces. The fact that Barno had not requested permission for this change and that the Pentagon had to learn about it from a newspaper also did not sit well with the leadership. Rumsfeld severely increased oversight on Barno, who now had to have weekly videocalls with Washington, in which he stated exactly how many Afghani forces he had trained. Simultaneously Rumsfeld ordered the Pentagon to develop its own new strategy for Afghanistan. When Barno was rotated out in 2005 by Gen. Karl Eikenberry, his Five Pillar plan was basically scrapped and focus returned on counterterrorism and training the ANA.47_{After his command in Afghanistan, Barno was}

appointed to a less-than-glamorous clerical position. He retired soon after and joined a think tank led by John Hagl, who was by then advocating the COIN approach throughout Washington.48

The Barno episode should perhaps not be seen as a completely isolated incident. The question of how to fight wars had actually been debated within Washington for quite some time. The military had in fact been ordered by Rumsfeld to develop new doctrine which emphasized the importance of stability operations, making it just as important as offensive and defensive military operations. This move encountered so much resistance that Rumsfeld appointed an outside manager to make sure it was implemented. This new doctrine, known as DoD Directive 3000.5, heavily accentuated stability operations, and would lay the groundwork for future publications on COIN operations.49

46_{Fred M. Kaplan, The Insurgents: David Petraeus and the Plot to Change the American Way of War (New}

York: Simon & Schuster, 2014), 333-334 and David W. Barno, “Fighting ‘The Other War’ Counterinsurgency Strategy in Afghanistan, 2003-2005,” Military Review 87, no. 5 (2007): pp. 32-44.

47_{Auerswald and Saideman, NATO in Afghanistan, 92-95 and Lebovic, Planning to Fail, 136-137.} 48_{Kaplan, The Insurgents, 334.}

49_{Benjamin Jensen, “Escaping the Iron Cage: the Institutional Foundations of FM 3-24. Counterinsurgency}

Doctrine,” Journal of Strategic Studies 39, no. 2 (2015): pp. 213-230, 220 and Jennifer M. Taw, Mission

(27)

23

Congress was also increasingly pushing the Pentagon to create better instruments for maintaining peace after combat. While these discussions were mostly focussed on Iraq, they also applied to Afghanistan. It appeared as though Congress and Rumsfeld were unified in their push for stability operations, but reality in the field proved different. It is unknown exactly why, but Rumsfeld quite simply refused to apply these approaches to Iraq or Afghanistan. So, while he was one of the main drivers for the development of such doctrine, its implementation was never ordered by him.50 In fact, as we saw in the Barno case, it was downright banned from the field.

As public attention also increasingly turned towards Iraq, less is known about public support for the Afghanistan war, let alone specific approaches to the conflict. One of the few data points is that a Gallup poll in 2004 showed for the first time that twenty-five percent of Americans considered the war in Afghanistan to be a mistake. While still a considerable minority, it was a significant rise from the situation in 2002 when only six percent considered it a mistake.51_There

were some attempts by Democratic Congress members to impose anti-war legislation, but these attempts were hardly successful and not broadly supported. With Republican control in Congress, these initiatives did not attain any results. Adding to this, most were concerned with the war in Iraq.52

In regards to Avant’s theory, this period is similar to the previous one as it can be applied to explain why the military did not achieve its preferred policy. Again, as the principals remained united, the military was unable to push its preference. Though Congress began making requests for new options, their opposition was not really focussed on Afghanistan, nor was it broadly supported. Therefore, it is here not regarded as a divided principal. The military preference was again not conservative, as Barno’s plans did not take into account Rumsfeld’s wishes. While the development of new doctrines was encouraged by Rumsfeld, the principal’s preference in the field remained unchanged. Feaver’s expectations are also applicable to explain events. As the principal’s monitoring of the agent decreased it led the military agent to think he would not be punished. The military agent then shirked by changing the strategy without consulting his superiors. When Gen. Barno’s shirking was noticed, he was punished through increased

50_{Maria Ryan, “‘Full Spectrum Dominance’: Donald Rumsfeld, the Department of Defense, and US Irregular}

Warfare Strategy, 2001–2008,” Small Wars & Insurgencies 25, no. 1 (February 2014): pp. 41-68, 41-44.

51_{Gallup, “More Americans Now View Afghanistan War as a Mistake,” Gallup (Gallup, February 19, 2014),}

https://news.gallup.com/poll/167471/americans-view-afghanistan-war-mistake.aspx.

52_{Michael T. Heaney, “The Partisan Politics of Antiwar Legislation in Congress, 2001-2011,” The University of}

(28)

24

monitoring and a not-too-glamorous new job. After this punishment, the military acted far more according to Rumsfeld’s preferences, and thereby worked in accordance with the theory.

4.3. Analysis

In the 2001-2006 period, both theories appear to be applicable to the case, as the expectations that are formulated provide an explanation to why events transpired as they did. Avant’s theory, however, seems somewhat more exact in its implementation.

In the early stages of the war Avant’s theory can be applied since there is a clear case of a united principal, as evidenced by Congress’ near unanimous support for the president’s conduct. The expectation would therefore be that the military would not be able to push its own preferences, and would instead follow the civilian’s more radical instruction. As can be seen in the rejection of military plans, and the following acceptance of the military of CIA plans, this expectation is fully met.

Feaver’s theory does not apply in this early stage. There is no detectable punishment, and fear of punishment does not seem to have been a major factor based on previous confrontation between the military and the civilian leadership. When the military preference for the Afghanistan invasion was denied, the military, counter to Feaver’s expectation in case of little punishment, chose to work instead of shirk. There could have been other reasons for the military’s decision to not shirk when given their orders. It may be argued that the military and the civilian in fact had converging interests, as the military saw benefit in not staying in Afghanistan with many troops. However, the primary preference by the military of slowly building a sizable force to invade was not attained, after which the military worked without punishment. The expectation, therefore, does not explain events.

In the 2003-2006 period, the application of the theories changes little. Avant’s expectation again explains events in this period. The principal remained united, despite the start of congressional opposition of war execution by asking for more options in regard to maintaining peace. The military, through the actions of Gen. Barno, also did not suggest a conservative policy since he completely ignored Rumsfeld’s wishes. As such, following Avant’s expectations, the military was unable to attain its preferences in the face of a united principal.

Feaver’s theory works better in this period than it did in the previous one. There is a clear sign of a reduced expectation of punishment, since Barno never got much attention from Washington. He then shirked, by implementing his own strategy without consultation. When

(29)

25

this was found out in Washington, he was punished, and the military then started working again. This fully follows Feaver’s expectation.

Taken together, it appears that Feaver’s expectation applies only in the second period, when there is clear conflict between the civilian and military leadership. Avant’s theory, on the other hand, is more clearly applicable in the last period. In table form, the results look like this:

Period Does it follow Avant’s expectations? Does it follow Feaver’s expectations?

2001-2002

Yes, theory explains events No, theory does not explain events 2003-

2005

(30)

26

5. 2006-2010

5.1. 2006-2008: Bush’s Turn

While the Iraq war falls outside the scope of this thesis, events in that war soon had another impact on policy in Afghanistan. As the war in Iraq deteriorated, the public opinion of the war and the Bush administration’s handling of it plummeted, leading the Republicans to lose their majority in both houses of Congress during the 2006 midterm elections.53 Furthermore, the military voiced its most vocal criticism of Rumsfeld, when former officers called for his dismissal.54 Officially, however, the military still advised a continued strategy of slowly drawing down forces and getting out of Iraq.55 What followed was a large-scale discussion about the strategy in Iraq within the administration, the military, and the public forum at large. To shortly summarize the events, while President Bush was looking for new options, a group of military officers, most famously Gen. Petraeus, had been convening to work out a COIN approach for the US military.56_{In the end and after much discussion and bargaining, President}

Bush, his civilian staff, and military advisors agreed on a COIN approach for Iraq, which became known as the Surge.57

After heavy fighting, the violence in Iraq seemed to diminish, apparently proving the surge approach successful. What exactly led to this success is hotly debated, but for this thesis it is important that for the actors at the time, this outcome in Iraq vindicated the decision to surge and apply the COIN approach. Meanwhile in Afghanistan, the developments in doctrine and approach had not yet made any significant changes in strategy. Commander in Afghanistan General Eikenberry had executed his orders from 2005, by following the train-and-withdraw strategy favoured by Rumsfeld until the general’s succession in 2007. In fact, despite the published Directive that proclaimed stability operations as just as important as combat operations, the Afghanistan strategy did not give it any importance.58

Three things did change in this period, however. First was the security situation itself. From 2005 onwards, the Taliban resurgence in the country was undeniable and taking a heavy toll on

53_{Stephen B. Dyson, “George W. Bush, the Surge, and Presidential Leadership,” Political Science Quarterly}

125, no. 4 (2010): pp. 557-585, 567.

54_{David S. Cloud, Eric Schmidt, and Thom Shanker, “Rumsfeld Faces Growing Revolt by Retired Generals,”}

The New York Times, April 13, 2006,

https://www.nytimes.com/2006/04/13/washington/rumsfeld-faces-growing-revolt-by-retired-generals.html.

55_{Peter D. Feaver, “The Right to Be Right: Civil-Military Relations and the Iraq Surge Decision,” International}

Security 35, no. 4 (2011): pp. 87-125, 104..

56_{Jensen, “Escaping the Iron Cage,” 218-225.}

57_{Feaver, “The Right to Be Right,” 103-104 and 106-107.} 58_{Auerswald and Saideman, NATO in Afghanistan, 93-94.}

(31)

27

coalition troops, Afghan security forces, and local civilians. Rumsfeld remained convinced, however, that the US would not need to increase its presence, since NATO and other allies would take over security tasks in the country.59 Second was the 2006 midterm election and its effects. Democrats secured a majority in both houses, and started an increase in anti-war rhetoric and legislation. Though it should be noted that this mostly concerned the Iraq war.60 Rumsfeld resigned in 2006, after this midterm election and the changing wind in Washington that followed. His last task as Secretary of Defense was to appoint the last commander in Afghanistan he approved of: General McNeill. The third thing to change was the Bush administration’s goals and concurrent strategic thinking on Afghanistan. A strategy review that started in 2006 and was published in 2007, but was mostly overshadowed by events in Iraq, stated that the US would now also aim at improving local governance, agriculture, and other civic and economic factors. Such goals are mostly linked to COIN or stability operations, and were in any case a clear break from previous policy.61

The change does not seem to have stemmed from a shift in public opinion on the war in Afghanistan, since this barely changed in this time frame. Although it must be mentioned that this can partly be explained by the lack of public attention to the conflict in this period. Perhaps unsurprisingly, congressional attention was also focussed on Iraq.62 Nevertheless, the new Secretary of Defense Robert Gates quickly set out to make sure this strategy would be put in the field in Afghanistan. Firstly, 3,200 extra Marines were deployed to help with the security situation. Secondly, General McNeill was replaced after just one year by General McKiernan, for failing to embrace the new strategic objective. And thirdly, the Bush administration appointed General Petraeus, of Iraqi Surge fame, as commander of CENTCOM. This command was responsible for the entire region, and Petraeus would therefore be McKiernan’s direct superior. Petraeus then announced a strategic reassessment of his entire command region, thus including Afghanistan.63

While the results of this review would only be available to Bush’s successor Barack Obama, it was to be expected that a COIN approach would be recommended in Afghanistan. While not actually locking in the next president, Bush had made it quite the undertaking for Obama to

59_{Auerswald and Saideman, NATO in Afghanistan, 95-97.} 60_{Heaney, “Partisan Politics of Antiwar Legislation”, 144-157.} 61_{Auerswald and Saideman, NATO in Afghanistan, 95-97.}

62_{Chris Good, “When and Why Did Americans Turn Against the War in Afghanistan?,” The Atlantic (Atlantic}

Media Company, June 22, 2011), https://www.theatlantic.com/politics/archive/2011/06/when-and-why-did-americans-turn-against-the-war-in-afghanistan/240880/.

Last War’s Theories in a New Conflict: Testing Civil-Military Relations Theories on the Afghanistan War (2001-2018)