American Civil–Military Relations and Presidential Power of Removal

• Published date: 01 July 2023
• DOI: http://doi.org/10.1177/0095327X221100270
• Author: Neil Snyder
• Date: 01 July 2023
• Subject Matter: Articles

https://doi.org/10.1177/0095327X221100270

Armed Forces & Society

2023, Vol. 49(3) 559 –592

© The Author(s) 2022

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0095327X221100270

journals.sagepub.com/home/afs

Article

American Civil–Military

Relations and Presidential

Power of Removal

Neil Snyder1

Abstract

Under what conditions are presidents more likely to remove senior defense officials

from office? Using a novel data set of all senior U.S. defense officials, both civilian

and uniformed military from 1947 to 2021, this article explores whether anticipated

support in Congress, partisan factors, or institutional protections for the military

affect observed patterns of removal for defense officials. The results suggest that

presidents are more likely to remove their own appointees (or their co-partisans’

appointees), but provide little evidence that presidents premise removal on

anticipated partisan support for a replacement nominee in Congress. Moreover, the

results suggest that military officials may have some insulation from politicization by

strong forms of removal, though that protection may weaken an important aspect

of civilian control of the military, the threat of punishment by removal, raising

profound questions for how civilian control endures.

Keywords

civil–military relations, political science, democracy, professionalism/leadership

During recent presidential administrations, there have been a number of widely pub-

licized removals of American defense officials, including both civilian and uni-

formed military officials. A few noteworthy examples include President George W.

1U.S. Department of the Army, Washington, DC, USA

Corresponding Author:

Neil Snyder, U.S. Department of the Army, 10053 East Legends Way, Fort Hood, TX 76544, USA.

Email: nnsnyder@gmail.com

1100270AFSXXX10.1177/0095327X221100270Armed Forces & SocietySnyder

research-article2022

560 Armed Forces & Society 49(3)

Bush’s removal of both his Secretary of Defense and the Chairman of the Joint Chiefs

of Staff during the Iraq war (Roberts, 2007; Spiegel, 2007; Stolberg & Rutenberg,

2006; Stolberg & Rutenburg, 2006), General Stanley McChrystal’s forced resigna-

tion under President Obama (“US Afghan Commander Stanley McChrystal Fired by

Obama,” 2010), and, in perhaps the most unusual instance of the period, President

Trump’s dismissal of Secretary of Defense Mark Esper via tweet (Detsch, 2020).

Although there is a robust literature on the politics of appointment anchored by

Lewis’s (2008) seminal work on the subject, to date scholars have devoted surpris-

ingly little attention to the politics of removal, leaving an important question unan-

swered: Under what conditions do presidents remove senior defense officials?

This article describes a systematic examination of the historical record of senior

defense officials under the contemporary American national security structure, from

the National Security Act of 1947 until the end of the last complete presidential

administration in 2021. Its focus is on removal from office: Is removal a politically

strategic act conditioned on the partisan composition of Congress? Are there observ-

able patterns of partisanship to the removal of defense officials? Do institutional

protections result in observed differences between how defense civilians and their

uniformed military counterparts exit office?

The results presented here provide little support for the hypothesis that presidents

condition their removal of defense officials on anticipated support in Congress or on

anticipation of so-called Senate “gate-keeping” (Bonica et al., 2015), but that firing

is likely influenced by partisan factors for both civilian and military officials.

Although civilian and uniformed military officials are removed at comparable rates

overall, there are observed differences in how military officials are removed, sug-

gesting that institutional differences may offer protection against one form of politi-

cization of the military (politicization through the firing of senior military officials):

a sign that presidents may be reluctant to pay political costs associated with firing

senior military officials; thus, protection from firing can also be protection from the

politicization of the military that would follow a politically-contested firing.

This article starts with a short case to illustrate the politics at play in presidents’

use of their removal power. From there, it reviews what the existing scholarship sug-

gests about how and why presidents wield the power of removal and it presents

several informal hypotheses that follow from this literature. Next, it introduces a

novel dataset and the results of statistical tests for whether certain conditions make

removal more likely. The article concludes with a discussion of the implications of

these results for American civil–military relations and suggests areas for continued

research.

The Removal of Secretary of Defense Donald Rumsfeld

Before proceeding with a deep dive into the politics of defense removals, a quick

detour is warranted to motivate this study with a well-known case: President George

W. Bush’s removal of Secretary of Defense Donald Rumsfeld in 2006 at a

Snyder 561

particularly dangerous point in the Iraq war (“Bush Replaces Rumsfeld to Get ‘Fresh

Perspective,’” 2006; McFadden, 2021; Roberts, 2007; Stolberg & Rutenberg, 2006;

Stolberg & Rutenburg, 2006). In the fall of 2006, Bush and his Republican col-

leagues were facing mid-term elections. Bush was entering his last 2 years of office

and his administration was embroiled in both the war in Iraq and the war in

Afghanistan. And the war in Iraq had been going particularly badly, with nearly

3,000 U.S. service members having been killed by election day in November (“Iraq

War in Figures,” 2011; Operation Iraqi Freedom [OIF] Casualty Summary by Month,

n.d.). The election was widely viewed as a referendum on the administration’s

perceived failings in Iraq (Nagourney, 2006; Washington, 2006). Bush’s approval

rating, once the highest approval rating ever measured by Gallup (Bush Job Approval

Highest in Gallup History, 2001), had plummeted to below 40% by early November

(Presidential Approval Ratings George W. Bush, 2008). On election day, the

Republicans lost control of both the House and the Senate, a sweeping change of

government perceived as devastating loss for which President Bush acknowledged

responsibility (“Bush Takes Blame for GOP Election Losses,” 2006).

On election day in 2006, Secretary of Defense Donald Rumsfeld was overseeing

the administration’s Iraq war policy. Rumsfeld, a second-time Secretary of Defense

and long-time ally of then Vice President Dick Cheney (McFadden, 2021), had been

an original member of Bush’s team (Allen, 2000). Donald Rumsfeld, who had been

closely associated with the administration’s policies ahead of the 2006 elections, was

no stranger to controversy, having been involved in several disputes inside the

Pentagon and with other members of Bush’s cabinet (Herspring, 2005). Ultimately,

Bush chose to replace Rumsfeld in early December, 2006, and chart a new course for

the war in Iraq (Bush, 2010). Robert Gates would succeed Rumsfeld, oversee the

implementation of a new strategy for Iraq (the so-called “Surge”) and eventually

decline to renew Marine General Peter Pace for another term as Chairman of the

Joint Chiefs of Staff to avoid a contentious confirmation process in the Senate (Gates,

2014). The administration nominated Admiral Michael Mullen to replace Pace

(Spiegel, 2007; Stout, 2007), effectively completing a changing of the guard at the

Pentagon in the fallout of the mid-term election.

Rumsfeld’s removal raises several questions about the political conditions affect-

ing presidential decisions to exercise their removal powers. At first, Bush’s decision

to fire Rumsfeld appears counterintuitive: With an opposition-controlled Senate tak-

ing control soon after Rumsfeld’s removal, Bush might have anticipated difficulty

getting a preferred nominee confirmed by a Democrat-controlled Senate. Rumsfeld’s

case raises the question as to whether, in general, presidents calibrate removal deci-

sions based on their perceived ability to get a preferred candidate confirmed by the

Senate. Rumsfeld’s case might suggest that defense removals are not conditional on

anticipated support in the Senate, otherwise Bush might have removed Rumsfeld

earlier and under more politically favorable conditions. Second, in January, 2007,

Bush faced a Congress controlled by his partisan opponents. On one hand, we might

expect the president to avoid major personnel changes to avoid potentially costly