Article
The sound of silence: Power,
secrecy, and international
audiences in US military basing
negotiations
Conflict Management and Peace Science
2014, Vol. 31(4) 406–431
Ó The Author(s) 2013
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0738894213508695
cmp.sagepub.com
Jonathan N. Brown
Central Michigan University, USA
Abstract
Why do leaders draw attention to some cooperative security negotiations but shroud others in
secrecy? Previous scholarship focuses on leaders’ efforts to gain better terms of agreement either
by playing their cards close to the vest at the bargaining table or by leveraging/avoiding aroused
public opinion at home. Yet, in many cases, it is neither dyadic nor domestic political pressures
that motivate leaders’ decisions to publicly acknowledge or conceal the occurrence of talks. This
article suggests, instead, that third-party states often constitute the primary targets of official
secrecy and that a state’s international power position shapes its decision to conceal or acknowledge military cooperation by affecting the size and attentiveness of international audiences, the
types of assets it brings to the relationship and the benefits it seeks from cooperation. I test five
hypotheses about leaders’ use of secrecy and acknowledgment through a statistical analysis of an
original dataset on US overseas military basing negotiations. This analysis produces strong support for my argument.
Keywords
Audience costs, international negotiations, military basing, secrecy, security cooperation
The Destroyers–Bases Deal of 1941 granted the USA rights to construct military bases on
eight British transatlantic possessions in exchange for 50 warships. American and British
officials reached this agreement after two rounds of publicly acknowledged negotiations during August 1940 to March 1941.1 A little known fact is that these negotiations were preceded
by secret talks during June to August 1939 regarding base rights in Trinidad, Bermuda, and
St Lucia. So secretive were these early negotiations that nearly all of the official records were
burned at the urging of US Under-Secretary of State Sumner Welles (Baptiste, 1976). These
initial talks were not publicly acknowledged until March 1946.2 Why did American and
Corresponding author:
Jonathan N. Brown, Department of Political Science, Central Michigan University, 247 Anspach Hall, Mount Pleasant, MI
48859, USA.
Email: [email protected]
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
407
British officials pursue highly secretive negotiations in June to August 1939 but openly
acknowledge their talks in August 1940 to March 1941?
Existing scholarship suggests that leaders engaged in security negotiations employ official
secrecy to withhold sensitive information from each other or domestic audiences. Rationalist
accounts of dyadic bargaining suggest that leaders use secrecy to keep the other side guessing
about their exact reservation point or level of military capabilities (Arena and Wolford,
2012; Lai, 2004; Meirowitz and Sartori, 2008; Slantchev, 2010) or when avoiding the attention of domestic audiences will significantly increase the prospects of cooperation, lower the
risk of costly fighting or decrease the political fallout from a foreign policy failure (Baum,
2004; Brown and Marcum, 2011; Kurizaki, 2007; Leventoğlu and Tarar, 2005; Putnam,
1988: 445; Stasavage, 2004; Tarar and Leventoğlu, 2009). Realist scholars similarly argue
that leaders resort to secrecy or deception when they anticipate domestic opposition to
unpopular foreign policies (Mearshiemer, 2011; Ripsman, 2002: 58–59; Schuessler, 2010).
Liberal theorists share this view that leaders ‘‘maintain secrecy mainly for domestic reasons’’
(Lipson, 2003: 87).
Yet, in the aforementioned talks between the USA and the UK, assessing whether either
side played their cards close to the vest while squaring off at the negotiating table does not
shed light on their decision to conceal the occurrence of the June to August 1939 talks but
then publicly acknowledge the August 1940 to March 1941 negotiations. Moreover, across
these time periods, the domestic political situations facing American and British officials
were not dramatically different.3 These cases are puzzling for the two conventional perspectives because a focus on dyadic bargaining dynamics and domestic audiences neglects the
fact that public disclosure of pending cooperation also invokes the attention of other state
leaders who are not participants in the negotiations. With few exceptions, existing scholarship gives scant treatment to third-party states—both allies and adversaries—as targets of
official secrecy.4 A corollary of this neglect is the literature’s inattention to how a state’s
power position in the international system shapes its leader’s management of national security information.
In this article, I argue, first, that third-party states often constitute primary targets of
information control during cooperative security negotiations and, second, that leaders’ use
of secrecy is shaped principally by the international strategic context. Leaders face three
dilemmas when choosing to conceal or acknowledge cooperative military negotiations:
enhancing their bargaining position at the risk of reducing the likelihood of cooperation;
accruing (avoiding) the precedent-setting benefits (costs) of concessions; and signaling
strength to an adversary at the risk of provoking preventive measures or sacrificing the
advantages of surprise. I argue that a state’s power position in the international system
shapes its leaders’ assessment of these dilemmas—and thus the decision to go public or
private—during cooperative security negotiations in three ways: by affecting the size and
attentiveness of international audiences, the types of assets it brings to the negotiating table
and the benefits it seeks from military cooperation. In these ways, this article makes novel
theoretical contributions to scholarship on bargaining and security cooperation by extending
the analysis of secrecy beyond dyadic and domestic political factors.
I evaluate this argument empirically by studying leaders’ control of information during
negotiations for foreign military base rights, where the sending state seeks privileges to station its defense personnel at naval, air or ground-force installations on the receiving state’s
territory (Harkavy, 1989). I chose this substantive domain for two reasons. First, although
foreign basing is among the most controversial—and therefore secretive—forms of security
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
408
Conflict Management and Peace Science 31(4)
cooperation, leaders often publicly disclose the occurrence of negotiations. In particular,
during the period of 1939–1971, the USA engaged in 218 rounds of basing talks. American
officials publicly acknowledged 54 (or 24.7%) of these rounds and host-state officials
acknowledged 64 (or 29.3%). Second, whereas most work on secrecy focuses on conflict scenarios between adversaries, the study of basing negotiations allows for an analysis of information control in cooperative security relationships during wartime and peacetime. A
statistical analysis using an original dataset of US overseas basing negotiations produces
strong empirical support for my argument. This evidence suggests that great power hosts are
more likely to employ secrecy than other types of receiving states and that both sending- and
host-state leaders generally seek to conceal the occurrence of basing negotiations when disclosure to international audiences may reflect poorly on their state’s prestige or unnecessarily
provoke rival states to take preventive measures in opposition to the talks. Alternatively,
leaders seek to invoke the attention of international audiences when doing so may set favorable precedents for future negotiations with other potential cooperative partners or signal
strength to wartime opponents.5
The remainder of the article proceeds as follows. The first section explains the three general dilemmas that leaders face when choosing to publicly disclose or withhold acknowledgment of pending security cooperation. The second section develops a theory of how states’
power positions in the international system shape leaders’ assessment of these dilemmas—
and thus their decision to go public or private—during military basing negotiations. The
third section describes the research design for the empirical tests. The fourth section reports
the results of the statistical analysis and provides historical examples of the main findings.
The article concludes by discussing directions for future research.
Going public or private during cooperative security negotiations
Leaders engaged in cooperative security negotiations must decide whether to directly invoke
the attention of outside audiences (going public) or to pursue talks secretly without their
knowledge (going private). Thus, the choice of secrecy/going private captures leaders’ concealment of the occurrence as well as the content of negotiations rather than simply the latter
(Pozen, 2010). Leaders face three general dilemmas when choosing to conceal or acknowledge pending military cooperation.
Constraints, bargaining leverage and the prospects of cooperation
The first dilemma arises when a leader faces constraints that simultaneously enhance his bargaining leverage but decrease the probability of cooperation (Putnam, 1988: 445; Schelling,
1960: 22–28). Public statements that invoke the attention of outside audiences capable of
imposing costs on leaders constitute one such constraint (Brown, 2014; Fearon, 1994;
Sartori, 2005). When the attention of outside audiences generates a higher level of constraints for one of the negotiating parties, that leader’s bargaining position often improves,
although at the cost of inducing them to adopt uncompromising demands and thereby
shrinking the set of terms that both parties can openly accept. This bargaining inefficiency
creates an incentive for leaders to pursue talks secretly (Kurizaki, 2007; Leventoğlu and
Tarar, 2005; Stasavage, 2004; Tarar and Leventoğlu, 2009).
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
409
Precedents: the benefits of success and the costs of failure
The second dilemma is that the attentiveness of outside audiences to leaders’ foreign policy
behavior simultaneously increases the benefits of success and the costs of failure (Baum,
2004). In particular, leaders manage information to accrue (avoid) the precedent-setting benefits (costs) of concessions (Lebow, 1996: 86–87). As Paul Anderson (1981: 741–743) notes,
‘‘Governments care about the expectations other governments infer from their behavior
because expectations make a difference . Unlike public diplomacy, secret foreign policy
actions will be relatively unconstrained by the influence of . precedents on expectations’’.
Thus, leaders may go public to set favorable precedents for future negotiations with thirdparty states—namely, allies or other potential cooperative partners—when their bargaining
position is strong. Alternatively, when their bargaining position is weak, leaders have an incentive to pursue talks privately to minimize the precedent-setting nature of their concessions.
Deterrence/compellence, provocation, and surprise
The final dilemma confronting leaders engaged in cooperative military negotiations concerns
striking a balance between improving their own security situation without needlessly frightening others (Jervis, 1976) or prematurely tipping their hand (Axelrod, 1979). Specifically,
leaders face the potential trade-off between the deterrent or coercive benefits from publicly
signaling strength to an adversary and the risk of diminishing the military effectiveness of the
relationship by unnecessarily provoking an opponent to take preventive measures or sacrificing the tactical advantages of surprise (Lai, 2004; Ritter, 2004; Slantchev, 2010).
To sum up, leaders face three dilemmas when choosing to conceal or acknowledge cooperative military negotiations: enhancing their bargaining position at the risk of reducing the
likelihood of cooperation; accruing (avoiding) the precedent-setting benefits (costs) of concessions; and signaling strength to an adversary at the risk of provoking preventive measures
or sacrificing the advantages of surprise. In the next section, I argue that the international
strategic context and the scope of states’ national security interests are primary causal factors influencing leaders’ assessment of these dilemmas.
Power, secrecy and international audiences in basing negotiations
I develop a theory of how states’ power positions in the international system shape leaders’
management of information relative to third-party states during cooperative security negotiations.6 Its logic is rooted in three propositions from the realist perspective on international
politics. First, given the lack of a supranational authority with the power to protect states’
independence, leaders alone must promote their state’s national interests. Second, leaders
can never be certain about other states’ present and future intentions. Accordingly, ensuring
the security of their state from external threat and preserving the autonomy of their foreign
policy from external influence constitute basic goals. Third, a state’s national interests and
efforts to advance them are driven principally by its power position in the international system (Mearsheimer, 2001; Morgenthau, 1978; Walt, 1987; Waltz, 1979). I argue below that a
state’s power position also shapes its leader’s decision to go public or private during military
basing negotiations in three ways: by affecting the size and attentiveness of international
audiences, the types of assets it brings to the negotiating table, and the benefits it seeks from
military cooperation.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
410
Conflict Management and Peace Science 31(4)
The size and attentiveness of international audiences
A state’s power position shapes its leaders’ management of national security information by
affecting the size and attentiveness of international audiences. A great power’s foreign policy
choices are likely to receive more attention from a wider international audience, including
allies and adversaries, because its behavior affects a broader range of actors and interests.
Consequently, a great power’s pursuit of a controversial foreign policy is likely to generate a
higher level of international attention and thus possibly increased constraints from thirdparty states opposed to the pending relationship. For the same reason, the precedent-setting
effects of concessions—whether positive or negative—also are likely to be more expansive
for great powers, given the scope of their other international relationships.
Most basing relationships are characterized by an asymmetry of capabilities between the
sending and host states. The sending state is invariably a great power and often enjoys a preponderance of material capabilities relative to the receiving state. Some hosts, however, also
are great powers. The argument above suggests that leaders should be less inclined to
acknowledge basing negotiations when both sending and host states are great powers for
two reasons.7 First, military cooperation between two great powers holds significant international implications owing to their potential aggregate capabilities. Consequently, public
acknowledgment of pending great-power cooperation will attract the attention of a wider
international audience. Rivals, in particular, may perceive such cooperation as provocative
and therefore have incentive to actively oppose the talks. Alternatively, a second reason is
that the decision by one great power to host another great power’s military forces on its territory may be perceived by international audiences as a signal of the former’s diminishing
prestige or status as a major player. As Robert Gilpin (1981: 31) notes, ‘‘Prestige is the reputation for power’’ and therefore ‘‘refers primarily to the perceptions of other states with
respect to a state’s capacities and its ability and willingness to exercise its power’’. Thus,
according to Hans Morgenthau (1978: 91), given that they ‘‘must rely upon their own power
for the protection of their existence and power position, [great powers] can hardly neglect
the effect that a . loss of prestige will have upon their power position on the international
scene’’. This suggests the first hypothesis.
H1: Sending- and host-state leaders are less likely to go public when both states are great powers.
Assets, coercion and bargaining leverage
The second way a state’s power position shapes its leaders’ management of national security
information is by affecting the types of assets it brings to the negotiating table. The assets
states bring to a cooperative security relationship vary in their specificity, ranging from
aggregate resources that are ‘‘highly flexible and easily transferred from one application to
another’’ to issue-specific assets that are ‘‘highly specialized and difficult to redirect’’ (Lake,
1999: 54). Whereas great powers exert influence with their aggregate economic and military
capabilities, weaker states gain bargaining leverage by manipulating relationship-specific
assets needed by the stronger state (Habeeb, 1988).
Basing relationships are characterized by this type of asymmetry of assets, for a weak
state that controls access to strategically located territory can use the threat of territorial
denial to reduce the overall leverage that the sending state can wield during negotiations
(Pape, 2005: 36).8 Consequently, the sending state’s material predominance does not
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
411
necessarily translate into a clear bargaining advantage (Fox, 1959; Keohane, 1971), except in
cases where its capabilities carry an explicit threat of coercion. In particular, some basing
agreements are outgrowths of colonial relationships or military occupations imposed on vanquished enemies. In such cases of ‘‘imposed cooperation’’, where ‘‘the stronger party in a
relationship can force the other side to alter its policies’’ (Milner, 1992: 469), the host-state
leader’s contractive capacity is sufficiently impaired that she has little choice but to make
any desired concessions (Moak, 1963: 19–20). Sending-state leaders can attempt to use such
unequal negotiations to set favorable precedents for future talks with other states. This suggests a second hypothesis.
H2: Sending- and host-state leaders are more likely to go public when the host state was recently
subjugated by the sending state.
The benefits of security cooperation
The third way a state’s power position shapes its leaders’ management of information is by
affecting the benefits it seeks from security cooperation. A cooperative military relationship
can advance a state’s security (i.e. its ability to preserve a desired status quo) or autonomy
(i.e. its ability to pursue desired changes in the status quo) (Morrow 1991). Basing relationships involve the host state offering concessions (i.e. access privileges) that increase the sending state’s freedom of action by extending its ability to project power abroad. In return, the
host state can receive a variety of quid pro quo, although most seek the security benefits
from hosting a foreign military. This asymmetry of benefits, with the sending state receiving
autonomy and the host state security, shapes leaders’ management of information during
basing negotiations by affecting how they assess the dilemma between deterring/compelling
and provoking adversaries.
A state’s ability to quickly project power declines with distance, while the threat posed by
an adversary increases with proximity (Boulding, 1962: 229–230, 245–247; Bueno de Mesquita,
1981: 40–44). A primary purpose of foreign military basing is to enhance the sending state’s
ability to overcome the time–distance problem inherent to power projection by establishing a
forward presence near potential theaters of operation (O’Hanlon, 2009: 141–168). Foreign
basing relationships are perceived as security-enhancing by host states and provocative by
adversaries because they bring the sending state’s military capabilities into closer proximity to
both. As a great power seeking to enhance its autonomy, the sending state will be more
inclined to pursue a basing relationship secretly when it expects an adversary to have heightened incentive and ability to forcefully oppose particular negotiations. This is especially likely
in two situations. The first is when the potential host territory is located on the immediate periphery of the sending state’s adversary, for a proximate threat is more likely to prompt the rival
to forcefully oppose the negotiations. The second is when the sending state runs the risk of
overextending its military capabilities. Specifically, taking on simultaneous military commitments in multiple disputes increases the risk of spreading the sending state’s forces too thin
(Kennedy, 1987). Thus, when the sending state is already engaged in a war, it will be more
likely to pursue basing negotiations secretly for fear of triggering a response by an enemy
against the host or some other territory where it may not be prepared to resist.
In both situations, the host-state leader also is likely to prefer secret negotiations to avoid
provoking preventive diplomatic or military retaliation from the sending state’s adversary. A
host state’s decision to grant base rights to a great power may create or exacerbate a security
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
412
Conflict Management and Peace Science 31(4)
dilemma with third-party states opposed to the military relationship. This dilemma arises
because the host state’s efforts to increase its security can reduce third parties’ sense of security either by threatening to diminish their ability to defend themselves or by altering their
assessment of the host state’s motives (Glaser, 2010). Whether such a spiral develops hinges
on the certainty with which third parties know about the pending relationship, which depends
on whether the negotiations are officially acknowledged, particularly by the host state. This is
because the act of official acknowledgment can constrain the weaker or more vulnerable partner’s subsequent bargaining position by enabling third-party states to credibly scrutinize its
intentions for pursuing military cooperation. Alternatively, absent public acknowledgment,
third-party opposition can be ignored or at least plausibly denied. An important consequence
of pursuing talks secretly, then, is to increase the size of the weaker partner’s bargaining
range—which makes agreement more likely—by removing the need to clarify its defensive
intentions to third-party states and thus diminishing the salience of its security dilemma as an
immediate constraint (Brown, 2014). This suggests the third hypothesis.
H3: Sending- and host-state leaders are less likely to go public when (a) the potential host territory
shares a border with an adversary of the sending state or (b) the sending state is involved in an
interstate war.
Alternatively, a weaker state pursuing security from a basing relationship will be more
inclined to publicly acknowledge negotiations when it already faces an immediate external
threat that may be countered or diminished by hosting a foreign military presence. In this
case, the host-state leader has incentives to use a public basing relationship to signal strength
to her adversary, for any perceived increase in the host state’s capabilities may serve to compel its adversary to seek a negotiated settlement. The host state is especially likely to prefer
public negotiations when it already is engaged in active conflict, for the fear of prematurely
provoking its opponent is no longer operative. The sending-state leader also will be more
likely to pursue public negotiations in this situation, anticipating enhanced bargaining leverage and thus more concessions. This suggests the fourth hypothesis.
H4: Sending- and host-state leaders are more likely to go public when the host state is involved in
an interstate war that may be influenced by the basing relationship.
A state also may seek security or autonomy through a policy of neutrality in ongoing
conflicts (Wolfers, 1962: 217–229). Leaders will be more likely to pursue basing negotiations
secretly when either is publicly committed to neutrality for three reasons. First, from the
sending state’s perspective as a great power, publicly pursuing base rights on neutral territory may set an unfavorable precedent that motivates adversaries to seek their own footholds in sideline states. Second, when the sending state itself is neutral, it may pursue basing
negotiations secretly as a strategy for discreetly building up its forward presence and preserving the advantages of surprise in anticipation of future conflict. Third, from the host’s perspective as a weaker neutral state, publicly granting base rights to a great power may invite
unnecessary constraints and vulnerability from opposition by the sending state’s adversaries.
This suggests a final hypothesis.
H5: Sending- and host-state leaders are less likely to go public when either state is publicly committed to a policy of neutrality in an ongoing conflict.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
413
Research design
I have gathered comprehensive data on US negotiations for base rights on foreign territory
where it did not already legally possess such privileges during 1939–1971.9 As indicated in
Table 1, the USA issued 90 proposals to open basing talks, of which 82 were accepted by
the potential host. These acceptances led to 218 rounds of bilateral talks. I identified this
population of cases through archival research at the National Archives and Records
Administration (College Park, MD, USA) and the Center for Military History (Fort
McNair, Washington, DC, USA), supplemented with published document collections and
secondary historical sources.
It is useful to highlight two scope conditions attendant to this data. First, the empirical
analysis focuses exclusively on US basing negotiations, even though other major powers
similarly have pursued their own foreign basing networks. While this limited focus reflects
purely practical constraints on systematic and comprehensive data gathering, it does nevertheless narrow the immediate applicability of the statistical results to US cases. However,
the centrality of the USA to the contemporary international system makes these cases particularly noteworthy and thus ‘‘not simply a parochial concern’’ (Fordham, 2008: 740). At the
same time, there is anecdotal evidence for other sending states that appears consistent with
my theoretical expectations and therefore suggestive of the generalizability of the empirical
findings to non-US cases. Some of this supplementary evidence is discussed in the
Conclusion. Second, the temporal bounds of my empirical analysis are confined to two eras
in which international strategic concerns and competition were heightened: the Second
World War and the first half of the Cold War. I chose this time period because the USA
began to expand its foreign basing network in 1939 and official records on cases from later,
less competitive phases of the Cold War are not widely accessible owing to the minimum 30year closure period for classified national security documents. This restriction points to
another important scope condition: the international motivation theorized to underpin leaders’ decisions about going public or private during basing negotiations may be especially
salient in periods of heightened strategic competition. The quantitative analyses partially
address this concern by encompassing variables that shape the intensity of international strategic competition, such as involvement in war and proximity to adversaries. While these
scope conditions certainly should be kept in mind when weighing the evidence presented
below, they also should be seen as fruitful avenues for future research on other sending
states and their cooperative partners.
The unit of analysis for the quantitative tests is the negotiation round rather than the
standard annual observation because more than one round of talks can take place in a single
year, and any given round of talks can spill over from one year to the next. By examining
the actual episodes of interest rather than yearly observations, this research design is better
suited to capture patterns of activity and substantive decisions that do not necessarily unfold
on an annual basis (Huth and Allee, 2002: 20–26).
The dependent variable and statistical model
Going Public. The dependent variable for both the sending and host states is dichotomous,
coded 1 if officials from that state publicly acknowledge a round of negotiations prior to its
commencement or while it is taking place. The code of 0 captures secret or unacknowledged
negotiations.10 The data were gathered through archival research—in the General Records
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
414
Conflict Management and Peace Science 31(4)
Table 1. The population of US overseas basing negotiations, 1939–1971
Host territory/state
US proposal date
Negotiation time period
Number
of rounds
Any rounds
acknowledged?
Brazil
Bermuda, St Lucia and
Trinidad/ UK
Fakarava Island/France
May 1939
June 1939
4
2
Yes
No
1
No
Mexico
Galapagos Islands/
Ecuador
Panama
Cuba
August 1939
October 1939
5
9
Yes
Yes
November 1939
May 1940
2
4
Yes
Yes
Dominican Republic
May 1940
2
No
Haiti
May 1940
2
No
Venezuela
Uruguay
May 1940
May 1940
3
2
No
No
Peru
May 1940
3
No
Colombia
May 1940
4
No
Ecuador
May 1940
2
No
Chile
May 1940
1
No
Costa Rica
June 1940
2
No
Nicaragua
June 1940
4
No
Honduras
June 1940
3
No
Guatemala
June 1940
3
Yes
El Salvador
Caribbean and North
Atlantic Territories/UK
Canada
June 1940
August 1940
1
2
No
Yes
August 1940
4
No
Paraguay
August 1940
1
No
Bolivia
August 1940
2
No
UK
Greenland/Denmark
Liberia
Iceland
Surinam/Netherlands
January 1941
February 1941
June 1941
June 1941
September 1941
3
1
2
2
1
No
No
No
No
No
Fiji/UK
October 1941
May 1939 to May 1942
June 1939 to November
1939
August 1939 to
November 1939
July 1940 to January 1943
November 1939 to March
1947
July 1940 to March 1942
August 1940 to
September 1942
August 1940 to
December 1941
August 1940 to
December 1941
August 1940 to July 1942
June 1940 to October
1940
September 1940 to July
1942
September 1940 to
September 1942
September 1940 to
January 1942
August 1940 to
September 1940
August 1940 to January
1942
August 1940 to
November 1942
September 1940 to
September 1942
September 1940 to
November 1942
September 1940
August 1940 to March
1941
August 1940 to February
1944
August 1940 to
September 1940
September 1940 to April
1942
January 1941 to July 1942
April 1941
June 1941 to March 1942
June 1941 to May 1942
September 1941 to
November 1941
October 1941 to August
1942
2
No
August 1939
(continued)
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
415
Table 1. (Continued)
Host territory/state
US proposal date
Negotiation time period
Number
of rounds
Any rounds
acknowledged?
Dutch East Indies/
Netherlands
Phoenix Islands/UK
October 1941
1
No
October 1941
3
No
Australia
October 1941
2
Yes
Line Islands/UK
October 1941
3
No
Western Samoa/New
Zealand
Solomon Islands/UK
October 1941
2
No
October 1941
2
No
New Caledonia/France
November 1941
2
No
New Hebrides/UK
November 1941
5
No
Dutch Antilles/
Netherlands
Bora Bora/France
December 1941
October 1941 to
November 1941
October 1941 to April
1947
October 1941 to May
1942
October 1941 to April
1947
October 1941 to March
1942
October 1941 to May
1946
December 1941 to March
1942
December 1941 to May
1946
December 1941 to
January 1942
December 1941 to
February 1942
January 1942 to May 1946
February 1942 to
September 1942
May 1942 to August 1942
August 1942 to
November 1942
November 1942 to March
1945
November 1942
1
No
2
No
3
3
No
No
2
1
No
No
3
Yes
1
Yes
2
Yes
2
1
No
No
4
No
3
No
1
No
3
1
2
Yes
No
No
2
No
4
Yes
3
Yes
December 1941
Ascension Island/UK
French Equatorial Africa/
France
Tongatabu/UK
Oman
January 1942
February 1942
February 1942
July 1942
Iran
October 1942
French North Africa/
France
French West Africa/
France
French Guiana/France
Ukraine/Soviet Union
November 1942
March 1943
October 1943
Azores/Portugal
November 1943
Eastern Siberia/Soviet
Union
Yugoslavia
November 1943
Philippines
Saudi Arabia
Ellice Islands/UK
April 1945
May 1945
November 1945
Gilbert Islands/UK
November 1945
Admiralty Islands/
Australia
Newfoundland/Canada
February 1946
December 1942
December 1944
October 1946
December 1942 to
February 1944
March 1943 to April 1943
February 1944 to March
1944
November 1943 to July
1945
June 1944 to April 1945
December 1944 to
January 1945
May 1945 to March 1947
May 1945 to August 1945
March 1946 to December
1946
March 1946 to December
1946
February 1946 to July
1947
November 1946 to March
1952
(continued)
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
416
Conflict Management and Peace Science 31(4)
Table 1. (Continued)
Host territory/state
US proposal date
Negotiation time period
Number
of rounds
Any rounds
acknowledged?
Ethiopia
Libya
November 1948
June 1950
3
6
No
Yes
Federal Republic of
Germany
Iceland
September 1950
March 1952 to May 1953
August 1951 to
September 1954
January 1951 to May 1952
3
Yes
1
No
Japan
February 1951
2
Yes
France
March 1951
2
Yes
Norway
April 1951
2
No
Spain
June 1951
4
Yes
Denmark
January 1952
2
Yes
Turkey
September 1952
2
No
Italy
December 1952
4
No
Greece
April 1953
1
No
Republic of Korea
Netherlands
August 1953
February 1954
7
3
Yes
Yes
Thailand
May 1954
2
Yes
Republic of China
(Taiwan)
Morocco
February 1955
February 1951 to May
1951
February 1951 to
February 1952
April 1951 to October
1952
August 1951 to October
1952
July 1951 to September
1953
September 1952 to
January 1953
February 1953 to June
1954
January 1953 to October
1954
August 1953 to October
1953
August 1953 to July 1966
February 1954 to August
1954
November 1966 to July
1968
August 1955 to July 1965
5
Yes
5
Yes
Pakistan
Federation of the West
Indies
Guyana
December 1957
December 1959
2
5
No
Yes
December 1959
2
Yes
Surinam/Netherlands
Australia
April 1960
September 1960
4
2
No
Yes
Indian Ocean Territories/
UK
Bahrain
October 1960
5
No
1
No
September 1950
June 1956
April 1971
May 1957 to December
1959
April 1958 to July 1959
December 1959 to
February 1961
September 1965 to May
1966
April 1960 to April 1962
September 1960 to May
1963
September 1962 to
December 1966
July 1971 to December
1971
Note: The eight rejected proposals were Argentina in October 1940, the Soviet Union regarding Siberia in January 1943,
Saudi Arabia in July 1944, the Soviet Union regarding the Kurile Islands in August 1945, Belgium regarding the Belgian
Congo in March 1948, Sri Lanka in September 1949, Pakistan in December 1955, and Bangladesh regarding Onn Kyunt
Island in March 1971.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
417
of the Department of State (RG 59) and the Records of the Foreign Service Posts of the
Department of State (RG 84) at NARA—and through supplementary research using published document collections, the ProQuest Historical Newspapers database, and secondary
historical sources.11
While both leaders separately decide whether to pursue public or private strategies in a
given round of talks, their individual choices are interrelated. Given dichotomous dependent
variables, a bivariate probit model is an appropriate statistical technique for capturing the
interrelated decisions of two separate actors. This model estimates a separate equation for
the likelihood that the sending- and host-state leaders will individually choose to go public
in a round of negotiations. The model also accounts for the interrelatedness of their choices
by incorporating the correlation of the errors between the two equations, which is expressed
as the parameter rho (r). A statistically significant and positive rho indicates that the unmeasured factors that affect the sending state’s equation also affect the host state’s equation in a
similar way. A significant and negative rho, on the other hand, indicates that the common
unmeasured factors affect the two equations in an opposite manner (Greene, 1997: 906–911;
Huth and Allee, 2002: 189–191; Huth et al., 2012: 24; Zorn, 2002: 156–157). I estimate a
bivariate probit model to examine the sending- and host-state leaders’ decisions to go public
during each negotiation round.
Since a leader’s decision to go public in any given round may be related to her decision to
enter into that round in the first place, I also test for possible selection bias by estimating a
Heckman or censored probit model. I found no evidence of selection effects biasing the analyses, and the results remained consistent.12 Consequently, I report only the bivariate probit
results below.
The independent variables
Where possible, the independent variables are measured on a month-specific basis to more
accurately capture their effects on leaders’ decisions to go public or private.13 When either
leader goes public, the data for the independent variables come from the month when this
decision is made. If neither leader goes public, the data for the variables are drawn from the
last month of that round of talks.
Great Power. This indicator is dichotomous, coded 1 if the host state was a major power
(Danilovic, 2002: 26–46, 225–230). The USA consistently was a major power.
Subjugation. This indicator is dichotomous, coded 1 if the host state was subjugated by the
sending state in the recent past—that is, defeated in war or occupied by the USA or was a
US colony within the past five years. Data were gathered from secondary historical sources
and the Correlates of War (COW) Wars dataset (Sarkees and Wayman, 2010).
Adversary Border. This indicator is dichotomous, coded 1 if the host territory shares a land or
sea border with a third state that is engaged in a militarized dispute with—or is a strategic
rival of—the sending state at the time of the negotiations. Data on borders come from the
COW Direct Contiguity and Colonial/Dependency Contiguity datasets (Stinnett et al.,
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
418
Conflict Management and Peace Science 31(4)
2002). Data on disputes and rivals come from the COW Militarized Interstate Dispute dataset (Jones et al., 1996) and Colaresi et al.’s (2007) strategic rivalries dataset.
Wartime. A separate variable is coded for the sending and host states. This indicator is
dichotomous, coded 1 if the state is engaged in an interstate war at the time of the negotiations that may be influenced by the basing relationship. Data come from the COW Wars
dataset.
Neutrality. This indicator is dichotomous, coded 1 if the sending or host state was publicly
committed to a policy of neutrality, non-belligerency or non-alignment in an ongoing conflict. The data were gathered from secondary historical sources.
Empirical analysis
The results of the bivariate probit analysis are presented in Table 2 and provide strong support for the hypotheses and the overall theoretical argument. Moreover, there is considerable
support for the use of a bivariate probit model. The rho parameter is positive and significant,
indicating that the sending- and host-state leaders’ decisions are highly correlated and that
the unmeasured factors affecting the sending state’s equation also affect the host state’s equation in a similar way.14 To aid interpretation of substantive effects, Table 3 reports the predicted probabilities for select significant variables.
The first hypothesis received strong support. Both sending- and host-state leaders are significantly less likely to go public during basing negotiations when both states are great powers.15 In particular, the USA is 49.2% less likely to go public and the host state is 76.1% less
likely. In August 1939, for instance, when the USA approached France for naval base rights
on Fakarava Island in the South Pacific, the following rationale was given for conducting
the negotiations secretly.
The [lease] of one of the islands of the Tuamotu Archipelago and the establishment of a naval base
there might be interpreted by Japan as part of an encircling movement by the only great power
which is at present in a position to wage effective warfare against Japan. This strategic threat to
Japan might be expected to render American–Japanese relations more difficult and more unfriendly
than they are at present and to increase the likelihood of eventual war between the two countries. It
could be further anticipated that . in the event of indecision on the part of Japan as to whether she
should join with Germany and Italy in a war against Great Britain and France, knowledge that the
United States was intending to establish a naval base in the Tuamotu Islands might influence Japan
into active participation in the war: Japan might conclude that it would be preferable to come to a
showdown militarily with Great Britain and France prior to development of the contemplated naval
base as such a base would weaken Japan’s position in the Pacific in the event of war with Great
Britain and France. (US Department of State, 1956: 529–530)
Alternatively, British officials often expressed the concern that publicly negotiating base
rights with the USA would be perceived internationally as ‘‘poor bargains’’ for the UK,
unbecoming of a great power, and therefore ‘‘threatening to the imperial status’’ (Leutze,
1977: 72–127; Whitham, 1996).
The second hypothesis also received strong support. Both sending- and host-state leaders
are significantly more likely to go public when the host state recently was defeated in war or
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
419
Table 2. Bivariate probit results for going public in a round of negotiations
Main international variables
Great Power
Subjugate
Adversary Border (US)
Wartime (US)
Wartime (Host)
Neutrality
Domestic and control variables
Large Coalition Size
Executive Constraints
Election (US)
Election (Host)
Press Coverage
Technical Round
Ally
Previous Public Round
Round
Constant
Sending state decision
to go public
Host state decision
to go public
20.784**
(0.386)
2.48***
(0.652)
21.32***
(0.382)
21.07***
(0.288)
1.31**
(0.546)
21.22***
(0.378)
21.01***
(0.379)
2.36***
(0.622)
21.14**
(0.465)
20.793***
(0.270)
1.03*
(0.625)
21.24***
(0.410)
22.70***
(0.629)
0.868*
(0.494)
0.002
(0.009)
0.002
(0.008)
1.52***
(0.356)
20.815**
(0.381)
20.087
(0.381)
0.677**
(0.332)
0.078
(0.102)
20.417
(0.536)
22.65***
(0.571)
1.21***
(0.446)
20.007
(0.011)
0.005
(0.008)
1.58***
(0.326)
20.279
(0.541)
0.082
(0.345)
0.929***
(0.331)
0.130
(0.097)
20.573
(0.508)
N = 218. Log pseudolikelihood = 289.77.
r = 0.983. Wald test of r: p = 0.000.
Clustered robust standard errors in parentheses.
***p \ 0.01, **p \ 0.05, *p \ 0.1.
occupied by the sending state or recently was a colony of the sending state. The USA is
67.7% more likely to go public when it has forcibly subjugated the host in the recent past.
The host state, in turn, is 164.8% more likely to go public. Thus, when the sending state’s
superior capabilities carry an explicit threat of force, it can attempt to utilize this leverage to
wrest more concessions through public negotiations that set favorable precedents for future
talks with other potential hosts. For instance, the USA publicly conducted negotiations with
the Philippines in 1946–1947 (a former colony) and Japan in 1951–1952 (a defeated and
occupied enemy) by effectively making favorable base rights a necessary condition for
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
420
Conflict Management and Peace Science 31(4)
Table 3. Impact of changes in variables on predicted probabilities for going public
International variables
Great Power
No (0)
Yes (1)
First difference
Percentage change
Subjugate
No (0)
Yes (1)
First difference
Percentage change
Adversary Border (US)
No (0)
Yes (1)
First difference
Percentage change
Wartime (US)
No (0)
Yes (1)
First difference
Percentage change
Wartime (Host)
No (1)
Yes (1)
First difference
Percentage change
Neutrality
No (0)
Yes (1)
First difference
Percentage change
Domestic variables
Large Coalition Size
No (0)
Yes (1)
First difference
Percentage change
Executive Constraints
Low (0)
High (1)
First difference
Percentage change
Press Coverage
No (0)
Yes (1)
First difference
Percentage change
Sending state
Host state
0.595
0.302
20.293
249.2%
0.364
0.087
20.277
276.1%
0.595
0.998
0.403
67.7%
0.364
0.964
0.6
164.8%
0.595
0.048
20.547
291.9%
0.364
0.097
20.267
273.3%
0.595
0.126
20.469
278.8%
0.364
0.120
20.244
267%
0.595
0.947
0.352
59.1%
0.364
0.745
0.381
104.6%
0.595
0.125
20.47
278.9%
0.364
0.062
20.302
282.9%
0.595
0.001
20.594
299.8%
0.364
0.002
20.362
299.4%
0.595
0.899
0.304
51.1%
0.364
0.773
0.409
112.3%
0.595
0.964
0.369
62%
0.364
0.891
0.527
144.7%
Note: The predicted probabilities represent the marginal probability of the sending or host state making concessions,
regardless of whether the other state also goes public. The baseline probability was calculated by setting all variables at
their mode (except for Election, which was set at its mean). The first difference was calculated by subtracting the
baseline predicted probability from the predicted probability following a discrete change in the variable of interest from
its low value to its high value, holding all other variables constant. The percentage change in probability was calculated
by dividing the discrete change in the predicted probability by the baseline probability and then multiplying by 100.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
421
granting both states independence (Berry, 1989: 13–68; Takemae, 2003: 503–506). It subsequently attempted to use these negotiations as a baseline during talks with South Korea and
Taiwan.16
The analysis demonstrated that, as suggested in H3(a), both sending- and host-state leaders are significantly less likely to go public during basing negotiations when the host territory
shares a border with an adversary of the sending state. The USA is 91.9% less likely to go
public and the host state is 73.3% less likely. US negotiations with Turkey in 1953–1954 constitute an example of this, given that the latter shared a land border with the Soviet Union.
The negotiators agreed that the ‘‘secrecy [of] military rights [talks was] of great importance
because of known Soviet sensitivity [to] actual US military operations [in] peripheral countries’’ and therefore necessary to ‘‘avoid [a] USSR reaction causing [them] to counterbalance
US operations by adjusting their own defense planning and estimates which tend [to] diminish the gains attendant [to] additional US bases’’.17 Moreover, as suggested in H3(b), both
the sending- and host-state leaders are significantly less likely to go public when the sending
state is engaged in an interstate war at the time of the negotiations. The USA is 78.8% less
likely to go public and the host state is 67% less likely. For example, despite general support
for a basing relationship in both countries, the USA and Ecuador held secret negotiations
for the Galapagos Islands in December 1941 to June 1942, for fear of triggering a Japanese
attack against Ecuadoran territory (Caribbean Defense Command, 1945).
As predicted by the fourth hypothesis, both sending- and host-state leaders are significantly more likely to go public when the host state is fighting an interstate war at the time of
the negotiations that may be influenced by the basing relationship. The USA is 59.1% more
likely to go public and the host state is 104.6% more likely. In January 1942, for instance, as
Japanese planes extended their operating radius to within 800 miles of Australia, having
recently bombed Australian airdromes in the Bismark Archipelago, it was announced that
Australian and American officials were conducting negotiations for naval base rights. Prime
Minister John Curtin emphasized the great boost to Australian defense capabilities from the
relationship (Bell, 1977: 28–29).18
Finally, the fifth hypothesis received strong support. Both sending- and host-state leaders
are significantly less likely to go public during basing negotiations when either leader is
openly committed to a neutral policy in an ongoing conflict. The USA is 78.9% less likely to
go public and the host state is 82.9% less likely. Given Portugal’s neutral status during
World War II, the USA’s negotiations for air base rights on Santa Maria Island in the
Azores during 1943–1945 were shrouded in secrecy for two reasons. First, Portuguese officials sought to avoid provoking German retaliation. President Antonio de Oliveira Salazar
emphasized that ‘‘the German Minister had been after him repeatedly with all sorts of questions about the possibility of the facilities in the Azores being used by the Americans; and
that German curiosity in this respect had been so keen that . the question may represent
the keystone of German policy with regard to Portugal’’. He would agree to negotiate only
‘‘as long as some formula were observed which would permit him . to deny officially to the
Germans that he had granted any special facilities to the United States in the islands’’.
Second, American officials feared that if ‘‘Portugal granted [bases in the Azores] to us as a
power not allied with Portugal . the Germans might cite this as a precedent for obtaining
similar concessions from Spain in the Balearics’’ (US Department of State, 1964: 571, 574–
575).
Several of the control variables for domestic political factors also generated significant
results. First, basing negotiations between two large-coalition (i.e. democratic) leaders are
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
422
Conflict Management and Peace Science 31(4)
significantly less likely to be public than negotiations between a sending-state leader with a
large coalition and a host-state leader with a small coalition. The USA is 99.8% less likely
to go public and the host state is 99.4% less likely when both have large domestic coalitions.
This finding is consistent with scholarship that suggests leaders are more likely to engage in
secret negotiations when public attention may generate similar levels of domestic audience
costs for both parties (Kurizaki, 2007; Leventoğlu and Tarar, 2005; Tarar and Leventoğlu,
2009). It also is supportive of recent work on large-coalition leaders’ enhanced incentive and
ability to secretly pursue controversial foreign policies (Brown and Marcum, 2011). Second,
sending- and host-state leaders are significantly more likely to go public when both leaders
face high levels of institutionalized constraints on their ability to determine policy and when
there is press speculation about the negotiations in either state. Specifically, the USA is
51.1% more likely to go public when the host-state leader also faces a high level of executive
constraints and 62% more likely following press speculation. Similarly, host-state leaders
are 112.3% more likely to go public in the face of high executive constraints and 144.7%
more likely following press speculation. These findings are consistent with conventional wisdom that institutional constraints such as legislative oversight and a free press generally
enhance the transparency of leaders’ foreign policy behavior. There is no evidence, however,
that electoral cycles shape leaders’ decisions to draw attention to international negotiations.
Overall, these additional findings indicate that domestic political institutions broadly affect
both sending- and host-state leaders’ management of national security information during
basing negotiations.
Conclusion
This article advances two main arguments about leaders’ control of information during
cooperative security negotiations. First, third-party states—both allies and adversaries—
often constitute primary targets of official secrecy. Second, leaders’ control of information is
shaped principally by the international strategic context and the scope of their states’
national security interests. Statistical analyses using an original dataset of US overseas
basing negotiations produce strong empirical support for my argument. This evidence suggests that great power hosts are more likely to employ secrecy than other types of receiving
states and that both sending- and host-state leaders generally seek to conceal the occurrence
of basing negotiations when disclosure to international audiences may reflect poorly on their
state’s prestige or unnecessarily provoke rivals to take preventive measures in opposition to
the talks. Alternatively, leaders commonly seek to invoke the attention of international audiences when doing so may set favorable precedents for future negotiations with other potential cooperative partners or signal strength to wartime opponents. In these ways, this article
makes novel contributions to scholarship on bargaining and security cooperation by extending the analysis of secrecy beyond dyadic and domestic political factors.
There are several avenues for future research. First, whereas this article explores the international strategic conditions under which leaders are more or less likely to withhold
acknowledgment of pending military cooperation, an equally important issue to consider is
how leaders’ management of information affects their bargaining behavior and the prospects
of achieving cooperation (Brown, 2014). In particular, are leaders more likely to make concessions or reach an agreement when they resort to secrecy or when they go public during
negotiations? A second direction for future research is to test the hypotheses in less
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
423
competitive international environments than the Second World War and the early Cold
War, given that the international motivation theorized to underpin leaders’ decisions about
going public or private may be especially salient during periods of heightened strategic concern. As suggested below, gathering data on other sending states with basing networks prior
to 1939 will contribute to this purpose. It also will be useful to examine cases of US basing
negotiations during the latter half of the Cold War as official records become available. A
third avenue for future research is to expand the theoretical and empirical analysis beyond
basing relationships to examine leaders’ use of secrecy in pursuit of other types of security
cooperation, such as alliances, coalition warfare, intelligence sharing, and the provision of
foreign military assistance. For instance, why were negotiations and arrangements for the
delivery of American arms to Yugoslavia shrouded in secrecy in 1949 but then openly
acknowledged in 1951 (Heuser, 1989: 162–164; Lees, 1997: 98–112; Lindsay, 1993: 337)?
Finally, future research could extend the empirical analysis beyond the USA. The arguments
developed above should be generalizable to other sending states and their respective host
partners. Testing this empirically will require systematic data gathering on other major powers that have pursued overseas basing networks. In the meantime, though, there is anecdotal
evidence for other sending states that appears consistent with my theoretical expectations. I
will conclude with a brief survey of these cases.
Consistent with the logic of H1, for instance, the secrecy surrounding Germany’s efforts
to secure base rights from Italy, a major power ally, in both Libya and the Dodecanese
Islands during August to November 1940 was doubtless related, at least in part, to international prestige issues: namely, Italy’s reluctance at having German forces employed in the
Mediterranean—its sphere of influence—and Mussolini’s increasing concern about being
perceived as Hitler’s vassal (Van Creveld, 1973: 52–61; Schreiber et al., 1995: 180–277). On
the other hand, in October 1968, two months after the Soviet suppression of the Prague
Spring, Alexander Dubček publicly was saddled with negotiating and signing an agreement
that allowed for the permanent stationing of Soviet troops in Czechoslovakia (Andrew and
Gordievsky, 1990: 487–488). Viewed in light of the emerging Brezhnev Doctrine and the
Soviet Union’s embarrassing loss of base rights in Albania a few years earlier, this case
would seem broadly consistent with the logic of H2: namely, the use of unequal negotiations
to secure major concessions that set favorable precedents for future talks with other host
states. In a similar vein, during August to September 1940, Japan capitalized on its ally
Germany’s recent subjugation of France to publicly pressure Vichy authorities in Indochina
to negotiate and sign an agreement for base rights in Tonkin (Hammer, 1954: 20–23).
Alternatively, the Soviet Union and Vietnam concealed their negotiations for base rights
at Cam Ranh Bay during 1977–1978 partly to avoid unnecessarily provoking China, a proximate rival to both states. This behavior seems consistent with H3. Moreover, as suggested
by the logic of H4, the situation changed in March 1979, one month after the outbreak of
the Sino-Vietnamese War, when Hanoi publicly acknowledged the relationship (Chin, 1986:
22–28; Gilks, 1992: 216–220, 228–231; Vogel, 2011: 274–276). Also in line with the logic of
H4, Syria openly acknowledged discussions and arrangements for an expanded Soviet military presence in 1980 as an effort to shore up its deteriorating security position relative to
Israel, Jordan and Lebanon (Ja’far, 1985: 265–268). The Soviets likewise capitalized on this
relationship as a useful counterpoint to the progressive and humiliating curtailment of their
access to Egyptian facilities (Harkavy, 1982: 187). Conversely, as suggested by the logic of
H5, the September 1940 talks between Germany and neutral Spain regarding base rights in
the Canary Islands were shrouded in secrecy (US Department of State, 1960: 90, 97, 100–
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
424
Conflict Management and Peace Science 31(4)
101, 107, 135, 168–171, 201–204). Likewise, early discussions in December 1961 for Soviet
base rights in Egypt were not publicly acknowledged since ‘‘President Nasser was firmly
wedded to the principles of non-alignment’’ (Dragnich, 1974: 22). Even after signing an
agreement in March 1968, Nasser continued to ‘‘publicly deny that Egypt had granted bases
to the Soviet Union’’ (Remmek, 1979: 372). These examples constitute suggestive evidence
not only of the generalizability of this article’s arguments and findings to non-US cases, but
also of the range of research opportunities open to scholars interested in secrecy and international security affairs.
Acknowledgments
I thank Austin Carson, Laura Dugan, Paul Huth, Scott Kastner, Tony Marcum, George Quester and
Jon Wilkenfeld for their feedback. I am also grateful to the editor of CMPS, Glenn Palmer, and three
anonymous reviewers for constructive input that helped strengthen the manuscript. An earlier draft
was presented at the 2011 International Studies Association conference in Montreal, Quebec, Canada.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-forprofit sectors.
Notes
1. ‘‘US is negotiating for British bases’’, New York Times, 17 August 1940, p. 1; ‘‘US delegation in
Britain to act on deal for bases’’, Chicago Daily Tribune, 25 January 1941, p. 3.
2. ‘‘Disclose secret accord of 1939 on island bases’’, Chicago Daily Tribune, 31 March 1946, p. 1.
3. President Franklin Roosevelt was no less constrained during the latter period by domestic political concerns about unpopular decisions like aiding Britain or expanding US naval activities in the
Atlantic than he was during the former. Neville Chamberlain’s hold on the premiership in the
summer of 1939 and the direction of his foreign policy were neither more nor less constrained by
domestic politics than were Winston Churchill’s a year later.
4. Ritter (2004: 36–81) examines secret alliances as a deterrence strategy, while Jervis (1997: 262) and
Morrow (2000: 70) suggest that states use secret alliances to avoid communicating limited commitments to outside states. Alternatively, Sartori (2005) argues that states use public diplomacy to
cultivate reputations for honesty in the eyes of international audiences.
5. This argument also sheds light on the puzzle of US–UK basing negotiations in the North Atlantic
and Caribbean during 1939–1941. Changes in the international strategic context after September
1939 altered both states’ (but especially the UK’s) incentives to signal strength to adversaries
rather than avoid provoking them, which was the case prior to that point. The USA also was presented with an opportunity to demand highly favorable terms of agreement that could be used as
precedents for future basing negotiations with other potential cooperative partners in the region.
Specifically, in the summer of 1939, Chamberlain was not keen to openly pursue Anglo-American
military cooperation for fear that it ‘‘might exacerbate the dictators’ sense of encirclement by the
democracies and make it harder to reach the European settlement that he so ardently desired’’
(Reynolds, 2006: 145). Conversely, Churchill remarked at a meeting of the War Cabinet in August
1940 that one public benefit of the Destroyers–Bases proposal was the ‘‘immense’’ impact it would
have on Germany, as the negotiations more closely associated the USA with the British war effort
(Leutze, 1977: 116–117). Alternatively, although Roosevelt still sought to avoid actions that might
precipitate US entry into the war, Britain’s desperate need for military aid in the autumn of 1940
presented a prime occasion to drive a hard bargain that favored US hemisphere defense plans
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
6.
7.
8.
9.
10.
11.
425
(Whitham, 1996). Indeed, Britain’s extensive concessions during the Destroyer–Bases negotiations
set a favorable baseline that US officials used in later talks with Latin American countries
(Weathers, 1960).
An important scope condition of my theory, then, is that it abstracts from dyadic bargaining and
domestic political factors. In doing so, I am not suggesting that these are unimportant. Rather,
the purpose of this theory is to rectify the neglect of international audiences—namely, third-party
states—as targets of secrecy in previous scholarship on cooperative security negotiations.
Since the sending state is always a great power, it is possible that the greater need for secrecy on
its part is a result simply of being the sender rather than necessarily a consequence of its great
power status. Of the 218 rounds of US basing negotiations during the period of 1939–1971,
American officials publicly acknowledged 54 (or 24.7%) and host-state officials acknowledged 64
(or 29.3%). The fact that the USA, as the sending state, goes public one-quarter of the time—and
this rate of acknowledgment is only slightly lower than the rate for host-states—suggests that simply being the sender does not, on its own, provoke a meaningfully greater need for secrecy.
The general point is that the host state always enjoys an advantage in issue-specific power relative
to the sending state because it controls access to the fixed asset of strategically located territory. Of
course, the specificity of any particular base or host state does vary depending on numerous contextual factors (see Cooley and Spruyt, 2009: 106). Consequently, one host state may possess more
issue-specific power relative to another host state. My theoretical argument, however, focuses on
the broader point about host states’ general advantage in issue-specific power relative to the sending state.
Renegotiation rounds for existing agreements are not included (see Cooley, 2008; Cooley and
Spruyt, 2009).
This dichotomous operationalization may strike some readers as problematic since some negotiations may have both public and secret aspects. However, this variable and the broader theoretical
argument capture something that is dichotomous: whether or not a state publically/officially
acknowledges the occurrence of negotiations. While talks may retain some secret aspects following acknowledgment, recognizing this does not weaken the argument. After all, it is the act of
acknowledgment that is the theoretical focus, not whether the subsequent negotiations are entirely
public.
For the coding of variables based on textual analysis, the issue of (intercoder) reliability becomes
a concern as more interpretation of the evidence is required. This hinges on two characteristics of
the variable being coded: distinguishability and observability. Less interpretation is needed as the
object being identified becomes more easily distinguishable within the body of evidence and when
the variable being coded is directly observable rather than latent (Rothman, 2007: 438–440). Thus,
descriptive objects generally have high levels of distinguishability and observability, especially
when compared to inferential objects that capture ‘‘intentions and impacts that confer meaning’’
(Druckman, 2005: 258). My chosen operationalization for the public acknowledgment variable
has the benefit of being descriptive and directly observable, thus requiring little interpretation.
Specifically, determining whether a state publicly acknowledges the occurrence of negotiations is a
straightforward matter: officials either acknowledge that negotiations are taking place—by issuing
a public statement—or they do not. No interpretation of intentions or impact is required to identify this basic act of acknowledgment. Instead, evidence that a public acknowledgment did occur
is distinguishable and observable in terms of (a) the internal government documents related to the
decision to issue an acknowledgment, (b) the various drafts of the statement, and (c) the text of
the released statement. The same can be said about evidence that a public acknowledgment did
not occur, as this can be identified through (a) the internal government documents related to the
decision not to issue an acknowledgment and (b) the lack of a statement. So while it certainly is
true that ‘‘no phenomenon is so clear-cut that counting it does not require some level of interpretation’’ (Colaresi et al., 2007: 29), minimal interpretation is required to code this variable.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
426
Conflict Management and Peace Science 31(4)
12. The selection model’s estimated rho parameter, which indicates the correlation of the disturbances
across the equations, did not approach standard levels of statistical significance. The details and
results of the selection analysis will be posted in an appendix on the author’s website.
13. Appendix 1 provides coding rules for the control variables.
14. Previous studies of international negotiations that use a bivariate probit similarly report high r
coefficients, often exceeding 0.95 (Huth and Allee, 2002: 240, 248, 259). Nevertheless, I conducted
three robustness checks. In each case, though, the results remained largely consistent. There also
are strong reasons for regarding the bivariate probit as a more useful and appropriate technique
than these alternate specifications. The first check was a probit model with a dyadic dependent
variable, where a code of 1 was assigned when either state went public. Compared with this specification, the bivariate probit is a more informative and useful model because it can capture the
different effects that independent variables might have across each leader’s decision. The second
check was a pair of separate probit models for the sending and host states. Again, though, a
bivariate probit is a more appropriate technique than two separate probits precisely because the
large and statistically significant r coefficient indicates that the equations are linked. Finally, I
conducted a multinomial logit using a dependent variable with four values: both private, US public/host private, US private/host public, both public. Given that the sending and host states’ decisions are separate but rarely diverge, the bivariate probit remains a more accurate and
appropriate model than this alternative specification. The results of the robustness checks will be
posted in an appendix on the author’s website.
15. Given that the UK accounts for the majority of cases of great-power hosts in the dataset, one reasonable concern is that the US–UK relationship is driving these results for H1. To assess this concern, I ran a robustness check with a UK dummy variable, thereby effectively separating those
cases from the other great-power host cases (i.e. France in 1939 and 1951–1952 and the Soviet
Union in 1943–1945). In these analyses, the results for the Great Power variable in each equation
remained negative and statistically significant. The only differences were lower p-scores for the
coefficients (i.e. stronger significance levels). For the host equation, the coefficient for the new
UK variable was negative and significant; for the US equation, it was negative but not significant.
Taken together, these results strongly suggest that the findings for H1 are not driven by the US–
UK relationship.
16. On South Korea, see NARA, RG 59, CDF 1950–54, 711.56395B, Box 3200; CDF 1955–59,
711.56395B, Box 2919. On Taiwan, see NARA, RG 59, CDF 1955–59, 711.56393, Box 2917.
17. Telegram 1405 from Washington to Ankara, 31 May 1953, NARA, RG 84, Entry 2454-A (350/
56/13/4), 430.3 Military Bases Overseas–Turkey, Box 37.
18. ‘‘US warships to use bases in Australasia’’, Washington Post, 5 January 1942, p. 1.
19. This coding scheme introduces a small amount of bias in favor of 0, as this value is assigned to
instances of first negotiation rounds—that is, observations for which there were no prior rounds
to acknowledge. As the only alternative would be to drop the first negotiation round for each case
from the dataset, the present coding scheme is tolerated as the lesser of two evils.
References
Anderson PA (1981) Justifications and precedents as constraints in foreign policy decision-making.
American Ju1ournal of Political Science 25(4): 738–761.
Andrew C and Gordievsky O (1990) KGB: The Inside Story. New York: HarperCollins.
Arena P and Wolford S (2012) Arms, intelligence, and war. International Studies Quarterly 56(2):
351–365.
Axelrod R (1979) The rational timing of surprise. World Politics 31(2): 228–246.
Baptiste FA (1976) The British grant of air and naval facilities to the United States in Trinidad, St.
Lucia, and Bermuda in 1939 (June–December). Caribbean Studies 16(2): 5–43.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
427
Baum MA (2004) Going private: Public opinion, presidential rhetoric, and the domestic politics of
audience costs in U.S. foreign policy crises. Journal of Conflict Resolution 48(5): 603–631.
Bell R (1977) Unequal Allies: Australian–American Relations and the Pacific War. Carlton, Australia:
Melbourne University Press.
Berry WE (1989) U.S. Bases in the Philippines: The Evolution of the Special Relationship. Boulder, CO:
Westview Press.
Boulding KA (1962) Conflict and Defense: A General Theory. New York: Harper.
Brown JN (2014) Immovable positions: Public acknowledgement and bargaining in military basing
negotiations. Security Studies, in press.
Brown JN and Marcum AS (2011) Avoiding audience costs: Domestic political accountability and
concessions in crisis diplomacy. Security Studies 20(2): 141–170.
Bueno de Mesquita B (1981) The War Trap. New Haven, CT: Yale University Press.
Bueno de Mesquita B, Smith A, Siverson RM and Morrow JD (2003) The Logic of Political Survival.
Cambridge, MA: MIT Press.
Caribbean Defense Command (1945) Procurement, Occupation, and Use of Air Bases in the Galapagos
Islands and at Salinas, Ecuador. Historical manuscript file 8-2.8 BL 1945. Fort McNair,
Washington, DC: Center for Military History.
Chin KC (1986) China’s War with Vietnam, 1979: Issues, Decisions, and Implications. Stanford, CA:
Hoover Institutions Press.
Colaresi MP, Rasler K and Thompson WR (2007) Strategic Rivalries in World Politics: Position, Space
and Conflict Escalation. New York: Cambridge University Press.
Cooley A (2008) Base Politics: Democratic Change and the U.S. Military Overseas. Ithaca, NY: Cornell
University Press.
Cooley A and Spruyt H (2009) Contracting States: Sovereign Transfers in International Relations.
Princeton, NJ: Princeton University Press.
Danilovic V (2002) When the Stakes Are High: Deterrence and Conflict among Major Powers. Ann
Arbor, MI: University of Michigan Press.
Dragnich GS (1974) The Soviet Union’s Quest for Access to the Naval Facilities in Egypt Prior to the
June War of 1967. Arlington, VA: Center for Naval Analyses.
Druckman D (2005) Doing Research: Methods of Inquiry for Conflict Analysis. Thousand Oaks, CA:
Sage.
Fearon JD (1994) Domestic political audiences and the escalation of international disputes. American
Political Science Review 88(3): 577–592.
Fordham BO (2008) Power or plenty? Economic interests, security concerns, and American
Intervention. International Studies Quarterly 52(4): 737–758.
Fox AB (1959) The Power of Small States: Diplomacy in World War II. Chicago, IL: University of
Chicago Press.
Gilks A (1992) The Breakdown of the Sino-Vietnamese Alliance, 1970–1979. Berkeley, CA: University of
California Press.
Gilpin R (1981) War and Change in World Politics. New York: Cambridge University Press.
Glaser CL (2010) Rational Theory of International Politics. Princeton, NJ: Princeton University Press.
Goemans HE, Gleditsch KS and Chiozza G (2009) Introducing Archigos: A dataset of political
leaders. Journal of Peace Research 46(2): 269–283.
Greene WH (1997) Econometric Analysis, 3rd edn.Upper Saddle River, NJ: Prentice Hall.
Habeeb WM (1988) Power and Tactics in International Negotiation: How Weak Nations Bargain with
Strong Nations. Baltimore, MD: Johns Hopkins University Press.
Hammer EJ (1954) The Struggle for Indochina. Stanford, CA: Stanford University Press.
Harkavy RE (1982) Great Power Competition for Overseas Bases: The Geopolitics of Access Diplomacy.
New York: Pergamon Press.
Harkavy RE (1989) Bases Abroad: The Global Foreign Military Presence. Oxford: Oxford University
Press.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
428
Conflict Management and Peace Science 31(4)
Heuser B (1989) Western ‘Containment’ Policies in the Cold War: The Yugoslav Case, 1948–53. New
York: Routledge.
Huth PK and Allee TL (2002) The Democratic Peace and Territorial Conflict in the Twentieth Century.
New York: Cambridge University Press.
Huth PK, Croco S and Appel B (2012) Law and the use of force in world politics: The varied effects of
law on the exercise of military power in territorial disputes. International Studies Quarterly 55(1):
17–31.
Ja’far KM (1985) The Soviet Union in the Middle East: A case study of Syria. In Cassen R (ed.) Soviet
Interests in the Third World. London: Sage, pp. 255–283.
Jervis R (1976) Perception and Misperception in International Politics. Princeton, NJ: Princeton
University Press.
Jervis R (1997) System Effects: Complexity in Political and Social Life. Princeton, NJ: Princeton
University Press.
Jones DM, Bremer SA and Singer JD (1996) Militarized interstate disputes, 1816–1992: Rationale,
coding rules and empirical patterns. Conflict Management and Peace Science 15(2): 163–213.
Kennedy P (1987) The Rise and Fall of the Great Powers: Economic Change and Military Conflict from
1500 to 2000. New York: Vintage Books.
Keohane RO (1971) The big influence of small allies. Foreign Policy 2(Spring): 161–182.
Kurizaki S (2007) Efficient secrecy: Public versus private threats in crisis diplomacy. American Political
Science Review 101(3): 543–558.
Lai B (2004) The effects of different types of military mobilization on the outcome of international
crises. Journal of Conflict Resolution 48(2): 211–229.
Lake DA (1999) Entangling Relations: American Foreign Policy in Its Century. Princeton, NJ: Princeton
University Press.
Lebow RN (1996) The Art of Bargaining. Baltimore, MD: Johns Hopkins University Press.
Leeds BA, Ritter JM, Mitchell SM and Long AG (2002) Alliance treaty obligations and provisions,
1815–1944. International Interactions 28(3): 237–260.
Lees LM (1997) Keeping Tito Afloat: The United States, Yugoslavia, and the Cold War. University
Park, PA: Pennsylvania State University Press.
Leutze JR (1977) Bargaining for Supremacy: Anglo-American Naval Collaboration, 1937–1941. Chapel
Hill, NC: University of North Carolina Press.
Leventoğlu B and Tarar A (2005) Prenegotiation public commitment in domestic and international
bargaining. American Political Science Review 99(3): 419–433.
Lindsay F (1993) Beacons in the Night: With the OSS and Tito’s Partisans in Wartime Yugoslavia.
Stanford, CA: Stanford University Press.
Lipson C (2003) Reliable Partners: How Democracies Have Made a Separate Peace. Princeton, NJ:
Princeton University Press.
Marshall MG and Jaggers K (2007) Polity IV Project: Political Regime Characteristics and Transitions,
1800–2007. College Park, MD: University of Maryland.
Mearsheimer JJ (2001) The Tragedy of Great Power Politics. New York: Norton.
Mearsheimer JJ (2011) Why Leaders Lie: The Truth about Lying in International Politics. New York:
Oxford University Press.
Meirowitz A and Sartori AE (2008) Strategic uncertainty as a cause of war. Quarterly Journal of
Political Science 3(4): 327–352.
Milner H (1992) International theories of cooperation among nations: Strengths and weaknesses.
World Politics 44(3): 466–496.
Moak DW (1963) Host country pressures against United States strategic bases. MA Thesis,
Georgetown University, Washington, DC.
Moregenthau HJ (1978) Politics Among Nations: The Struggle for Power and Peace, 5th edn. New
York: Knopf.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
429
Morrow JD (1991) Alliances and asymmetry: An alternative to the capability aggregation model of
alliances. American Journal of Political Science 35(4): 904–933.
Morrow JD (2000) Alliances: Why write them down? Annual Review of Political Science 3: 63–83.
Morrow JD, Bueno de Mesquita B, Siverson RM and Smith A (2008) Retesting selectorate theory:
Separating the effects of W from other elements of democracy. American Political Science Review
102(3): 393–400.
O’Hanlon ME (2009) The Science of War: Defense Budgeting, Military Technology, Logistics, and
Combat Outcomes. Princeton, NJ: Princeton University Press.
Pape RA (2005) Soft balancing against the United States. International Security 30(1): 7–45.
Pozen DE (2010) Deep secrecy. Stanford Law Review 62(2): 257–340.
Putnam RD (1988) Diplomacy and domestic politics: The logic of two-level games. International
Organization 42(3): 427–460.
Remmek RB (1979) The politics of Soviet access to naval support facilities in the Mediterranean. In
Dismukes B and McConnell J (eds) Soviet Naval Diplomacy. New York: Pergamon Press, pp. 357–
403.
Reynolds D (2006) From World War to Cold War: Churchill, Roosevelt, and the International History of
the 1940s. New York: Oxford University Press.
Ripsman NM (2002) Peacemaking by Democracies: The Effect of State Autonomy on the Post-World
War Settlements. University Park, PA: Pennsylvania State University Press.
Ritter JM (2004) Silent partners and other essays on alliance politics. Ph.D. dissertation, Harvard
University, Cambridge, MA.
Rothman SB (2007) Understanding data quality through reliability: A comparison of data reliability
assessment in three international relations datasets. International Studies Review 9(3): 437–456.
Sarkees M and Wayman F (2010) Resort to War: A Data Guide to Inter-State, Extra-State, Intra-State,
and Non-State Wars, 1816–2007. Washington, DC: CQ Press.
Sartori AE (2005) Deterrence by Diplomacy. Princeton, NJ: Princeton University Press.
Schelling TC (1960) The Strategy of Conflict. Cambridge, MA: Harvard University Press.
Schrieber G, Stegemann B and Vogel D (1995) Germany and the Second World War, Vol. III: The
Mediterranean, South-east Europe, and North Africa 1939–1941. Oxford: Oxford University Press.
Schuessler JM (2010) The deception dividend: FDR’s undeclared war. International Security 34(4):
133–165.
Slantchev BL (2010) Feigning weakness. International Organization 64 (3):357–388.
Stasavage D (2004) Open-door or closed-door? Transparency in domestic and international bargaining.
International Organization 58(4): 667–703.
Stinnett DM, Tir J, Schafer P, Diehl PF and Gochman C (2002) The Correlates of War Project direct
contiguity data, version 3. Conflict Management and Peace Science 19(2): 58–66.
Takemae E (2003) The Allied Occupation of Japan. New York: Continuum.
Tarar A and Leventoğlu B (2009) Public commitment in crisis bargaining. International Studies
Quarterly 53(3): 817–839.
US Department of State (1956) Foreign Relations of the United States: 1939, Vol. II: The British
Commonwealth and Europe. Washington, DC: US Government Printing Office.
US Department of State (1960) Documents on German Foreign Policy 1918–1945: Series D (1937–
1945), Vol. XI: The War Years, September 1, 1940–January 31, 1941. Washington, DC: US
Government Printing Office.
US Department of State (1964) Foreign Relations of the United States: 1943, Vol. II: Europe.
Washington, DC: US Government Printing Office.
Van Creveld ML (1973) Hitler’s Strategy 1940–1941: The Balkan Clue. New York: Cambridge
University Press.
Vogel EF (2011) Deng Xiaoping and the Transformation of China. Cambridge, MA: Harvard University
Press.
Walt SM (1987) The Origins of Alliances. Ithaca, NY: Cornell University Press.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
430
Conflict Management and Peace Science 31(4)
Waltz KN (1979) Theory of International Politics. Boston, MA: McGraw Hill.
Weathers BE (1960) Acquisition of Air Bases in Latin America: June 1939–June 1943. Air Force
Historical Research Agency Historical Study No. 63.
Whitham C (1996) On dealing with gangsters: The limits of British ‘generosity’ in the leasing of bases
to the United States, 1940–41. Diplomacy and Statecraft 7(3): 589–630.
Wolfers A (1962) Discord and Collaboration: Essays on International Politics. Baltimore, MD: Johns
Hopkins University Press.
Zorn C (2002) U.S. government litigation strategies in the federal appellate courts. Political Research
Quarterly 55(1): 145–166.
Appendix 1: Measurements for control variables
Coalition Size—this variable is a control for domestic regime type. A dichotomous indicator
is constructed from Bruce Bueno de Mesquita et al.’s (2003) five-point winning coalition
scale, which ranges from 0 (the smallest coalition) to 1 (the largest coalition). To capture
large-coalition leaders, a code of 1 is assigned if the host-state leader’s coalition size equals
0.75 or 1. The code of 0 is assigned to leaders with small coalitions equal to 0, 0.25 and 0.5.
This variable is coded only for the host state, as the USA consistently had a large coalition.
I chose this operationalization of regime type for two reasons. First, previous scholarship
demonstrates that selectorate theory’s argument regarding coalition size and the provision
of goods is useful for explaining leaders’ behavior during basing negotiations (Cooley,
2008). Second, recent scholarship also links variation in coalition size to leaders’ incentive
and ability to engage in secret foreign policy behavior (Brown and Marcum, 2011).
Executive Constraints—this indicator measures the degree of institutionalized constraints
on the power of the executive to determine policy. It controls for the elements of democracy
outside of the winning coalition: namely, legislative constraints (Morrow et al., 2008). This
indicator is dichotomous, coded 1 if the EXCONST variable in the Polity IV dataset is equal
to 7, which captures executive parity or subordination to a relevant accountability group
such as a legislature (Marshall and Jaggers, 2007). This variable is coded only for the host
state, as the USA consistently had a value of 7 for EXCONST.
Election—this variable controls for cycles of political vulnerability, as previous scholarship suggests that leaders are significantly less likely to go public just prior to elections than
immediately after them (Leventoğlu and Tarar, 2005: 422). A separate variable is coded for
the sending and host states. This indicator is coded as the number of months since the last
relevant national election to account for endogenous election timing in parliamentary systems (Huth and Allee, 2002: 96). Information on the dates of all national elections comes
from the Archigos dataset (Goemans et al., 2009).
Press Coverage—this indicator is dichotomous, coded 1 if the national press in either the
sending or host state printed news stories reporting prospective US military basing in the
potential host territory within at least three months prior to the negotiation round. Data for
the USA were coded from press coverage in The New York Times, The Washington Post and
The Chicago Tribune, using the ProQuest Historical Newspapers database. Data for the
potential host territory came from press coverage in national newspapers in each potential
host state as comprehensively tracked by the US Department of State in RG 59 and RG 84
at NARA.
Technical Round—this indicator is dichotomous, coded 1 if the negotiation round is
devoted primarily to technical military details about the force structure deployed in the host
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Brown
431
state. Given that the disclosure of such information to an enemy could undermine established war plans or sacrifice the immediate tactical advantages of surprise, negotiation
rounds devoted primarily to technical military details are more likely to be shrouded in
secrecy. The data were gathered from archival sources.
Ally—this indicator controls for the effects of preexisting security cooperation. This indicator is dichotomous, coded 1 if the sending and host states share a military alliance. The primary source of data is the Alliance Treaty Obligations and Provisions dataset (Leeds et al.,
2002).
Previous Public Round—this indicator is dichotomous, coded 1 if the previous negotiation
round was public, as a control for past history.19
Round—this indicator counts the number of previous rounds, as one control for the nonindependence of observations.
Downloaded from cmp.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016