An opportunity cost theory of US treaty behavior*
Judith Kelley and Jon Pevehouse
WORKING PAPER: DO NOT CITE WITHOUT PERMISSION
Please do not cite without authors’ permission
jkelley@duke.edu
pevehous@polisci.wisc.edu
* We thank Kathryn Chylla, Frank Halloran at the US State Department, John Gamble
for sharing his data with us, and participants at the November 6, 2010 Duke-Harvard
Foreign Relations Workshop, "The Political Economy of U.S. Foreign Relations Law,”
Duke University. This paper was also presented at the Annual meeting of the American
Political Science Association, Seattle, WA, 2011, and the Annual meeting of the
International Studies Association, San Diego, April 1-4, 2012. For very helpful
comments we thank workshop participants at Georgetown University, Florida State
University, the University of Pennsylvania, and Cornell University. We also thank Mark
Pollack for additional comments.
1 An opportunity cost theory of US treaty behavior
What factors influence whether the US participates in global treaty-based cooperation?
Traditional explanations hold that states act based on their national interest and, for the
US this means that as long as two-thirds of senators are convinced that the benefits of the
treaty outweigh the costs, the US should participate in the treaty. However, at times the
US fails to join treaties that appear to have the support of the requisite 2/3 of senators, or
even quite uncontroversial treaties. This has long ago earned the senate a reputation as
“the graveyard of treaties.” This paper argues that this occurs because the
constitutional advice and consent process imposes opportunity costs for both the
president and the senate in terms of senate floor time, and that these costs slow down the
process significantly or even block it completely. A statistical analysis of all multilateral
treaties after 1967 shows that while, as expected, the traditional political and treatyspecific factors influence a treaty’s success in the senate, so do opportunity cost. This
means that the president and the senate do not simply consider each treaty on its own
merits; rather treaties may be sidelined due to the opportunity costs associated with
dealing with them.
The US Senate can be quite speedy at ratifying treaties. In March 2003, the US signed
the NATO protocols to enable the accession of several new members. By April the
president had sent the treaty to the senate and by May the senate provided its advice and
consent. However, such speed is rare. For example, also in 2003, the World Health
Organization concluded the Framework Convention on Tobacco Control (FCTC) and the
United States signed the treaty. Then Senator Barack Obama, a long time tobacco control
advocate also fighting personal tobacco addiction, joined in a letter urging President
George W. Bush to the send the FCTC to the senate for ratification. Now President
Barack Obama has now been president for four years. Why then has he failed to transmit
the treaty?
Similarly, in 1994, President Clinton transmitted the Law of the Sea and the Agreement
on Part XI to the Senate. Since then, the Law of the Sea has been on the Treaty Priority
List1 of Presidents Clinton, Bush, and Obama. Moreover, it has had the support of over
two-thirds of the Senate.2 Why then has it remained unperfected? Why do some treaties
remain on the president’s desk or remain tabled in the Senate despite strong incentives to
ratify the treaty?
1
Administrations have begun to issue “treaty priority list” for each new Congress. These typically list (1)
treaties currently on the Committee's calendar on which the Administration supports Senate action at this
time, (2) treaties currently on the Committee's calendar on which the Administration does not support
Senate action at this time, and (3) treaties not yet before the Committee on which the Administration is
actively reviewing treaty transmittal packages with a view toward Senate action prior to adjournment of the
Congress. For an example, see http://www.state.gov/documents/organization/116355.pdf and
http://globalsolutions.org/files/general/White_House_Priorities_List.pdf
2
http://www.nytimes.com/gwire/2009/05/07/07greenwire-sen-kerry-looks-for-window-to-ratify-law-of-th12208.html
2 In this paper, we formulate an answer to these questions. Specifically, we assess the
question: Once the United States has signed an international agreement that requires
Senate advice and consent, what explains how long the president takes to transmit it to
the Senate, and if he does transmit it, what explains how long the Senate takes to bring it
to a vote or whether the Senate ever gives it advice and consent at all? Indeed, the US is
often in the forefront of treaty creation and plays a formative role in the formulation of a
majority of treaties. Why then does the US lagged behind other states in treaty
ratification? One study of 76 treaties open to all states and concluded between 1990 and
2005 on a broad range of issue areas found that the US had only 34 treaties, far behind
European countries and the second lowest among Organization for Economic
Cooperation and Development countries (Elsig, Milewicz et al. 2012: 8).
Legal scholars have increasing pointed to the Senate’s failure to ratify treaties, yet
scholars have not explained why this occurs so often. Although there has been some
analysis of the Senate advice and consent process for presidential nominations (McCarty
and Razaghian 1999 ; Binder and Maltzman 2002 ; Derouen, Peake et al. 2005), scholars
have paid very little attention to the US treaty advice and consent process. We know of
only one broad and systematic examination of the US treaty process across treaties and
time (Krutz and Peake 2009), one study of determinants of treaty reservations attached by
the Senate (Auerswald and Maltzman 2003), and a few analyses of security treaties
(Krepon and Caldwell 1991 ; Delaet and Scott 2006), or case studies of individual treaties
(Evans and Oleszek 2003).3 There has been more attention to what determines whether
the US administration treats a particular international agreement as an executive
agreement or as a treaty that requires Senate advice and consent (Martin 2005 ; Krutz and
Peake 2009), than to the advice and consent process itself.
Yet, the Senate advice and consent process for treaties is important for several reasons.
First, the treaties that the administration determines requires senate advice and consent
tend to be, by constitutional design, matters of significant US commitments, sometimes
involving delegation of significant authority or resources. Second, whether the Senate
provides advice and consent, and whether the US ultimately ratifies treaties influences
how the US participates in international cooperation. For example, a lack of advice and
consent, and the subsequent delay in US treaty ratification prevents the US from
participating fully in the organizational work that flows from the treaty. Thus, the US
may be relegated to observer status, despite actively cooperating with the treaty
organization, a point the president sometimes bemoans in his transmittal letters to the
senate.4 Furthermore, given its global leadership role, US participation can be pivotal in
3
Note that there has been more emphasis on whether the president seeks to evade the senate advice and
consent process through use of executive agreements, a debate we set aside for now, but see our discussion
under data clarifying our sample identification process. (Martin 2005 ; Krutz and Peake 2009)
4
See for example President Bill Clinton’s Message Letter of Transmittal accompanying the South Pacific
Regional Environment Programme Agreement. Treaty Doc. 105-32, November 7, 1997, in which the
submittal letter from the State Department, which the president sends to the senate along with the
president’s own message of transmittal, notes that lack of US ratification will leave the United States
without influence on important committees within the treaty organization.
3 whether international cooperation on a particular issue succeeds. In cases such as the
Kyoto Protocol on global warming, the absence of Senate advice and consent may
diminish global efforts to cooperate on important issues, thus having important effects
not only for the US but for other states as well.
Lastly, during negotiations of a treaty, when other states surmise that the US will not
ultimately join the treaty regime, the US loses influence over the negotiations (Galbraith
2012). This in turn further diminishes the likelihood that the US will join the treaty and
thus bolsters even further skepticism about US good faith during treaty negotiation. A
former Legal Advisor to the State Department stressed this as a big problem for US
negotiators, because “It may well be that we will sign the treaty, but our negotiating
partners have no confidence that the executive branch will necessarily be able to get a
potentially controversial treaty through the Senate (Johnson 2010).” Note that this
dynamic also is contradictory to existing international relations theory: two-level games
literature traditionally suggests that domestic ratification difficulties can grant
international negotiators leverage in negotiation (Putnam 1988). Thus, treaties requiring
Senate advice and consent represent a particularly interesting body of international
agreements and it is important to understand how this domestic process influences the US
ability to cooperate with other nations.
Traditional national interest arguments
In the international relations theory, most existing analysis of treaty ratification is crossnational and treats states as rational, unitary actors. Indeed, nearly all of the analysis of
treaty ratification examines cross-national behavior on a given treaty. Thus, several
scholars compare decisions across states for specific treaties and demonstrate various
logics that drive state participation in treaties (Fredriksson 2000 ; Moravcsik 2000 ;
Simmons 2000 ; Buthe and Milner 2008 ; Simmons 2009). The logic may vary across
issues, but the underlying assumption usually is that states are unitary actors and that they
join treaty regimes that are in their national interest. Sometimes the benefit is a signaling
value that states attribute to joining a treaty even if they have no intent of observing it
(Vreeland 2008), but the fact remains that a state is thought to weigh the pros and cons of
the treaty and join if the benefits outweigh the costs. Based on this logic, if a treaty is in
the interest of the US, the US should ratify it. Thus, the US failure to join various treaties
such as the International Criminal Court is explained by the costs to the US of agreeing to
the treaty as exceeding the benefits. That is, existing theory essentially base state
decisions to join treaty regimes as only on how the attributes of the treaty serve the
national interest.
However, although this frame of analysis can take us a long way in understanding the fate
of many treaties, it leaves some puzzles, because there appear to be several cases where
the US clearly values a treaty, yet does not ratify it. For example, US negotiators
sometimes push hard for certain provisions but then, even when the US manages to
obtain nearly all its demands, the US fails to ratify the treaty. The American efforts to
renegotiate Article 11 of the Law of the Sea relating to deep sea bed mining are a case in
point. Sometimes the US even initiates treaties and shapes their form, and then fails to
4 ratify them. For example, the US has not joined the ILO Convention (185) on Seafarers’
Identity Documents, which was written in direct response to 9/11 terrorist attacks and
aimed to strengthen port of entry security. The US was the main driver behind this
convention and the tripartite US delegation voted unanimously to adopt it in 2003. But
the treaty has been held up in the Senate, supposedly due to a visa provision that runs
contrary to some federalism provisions in US immigration law.5 But why would
federalism trump national security? Thus, the strong negotiating behavior of the US in
this case signals that even treaties that are in the US interest can remain unperfected.
Similarly, if the US should become party to treaties that are in their interests, it is also a
puzzle why the US sometimes participates actively in some treaty regimes, seeking to
meet the requirements and build their organizational structures, but does not ratify the
underlying treaty. For example, the US played a major part in the treaty negations for the
Biodiversity Convention and has an excellent compliance record, but has not ratified the
treaty itself. Likewise, the US participates in research and development for the
Agreement on Conservation of Albatrosses and Petrels. President George Bush Jr.
transmitted the treaty to the senate in 2008, 7 years after its conclusion, endorsing it as
being fully in the US interest, yet the treaty still lingers in the Senate. Why does the
United States actively support and promote the use of international bodies but then fails
to ratify the treaty documents generating these bodies?
The US behaves in other ways that signal that it endorses some treaties that it nonetheless
has not ratified. In trade law and other law, the US will refer to unperfected treaties as
precedential for US-authored treaties. For example, although the US has not ratified most
ILO Conventions, it is reinforcing the content of many those unratified conventions in a
trade agreement with Peru (Charnovitz 2008 96).Why does the US incorporate into trade
agreements various human rights conditions (Hafner-Burton 2009), some of which the
US has not ratified and accepted into its own domestic law? This suggests that,
paradoxically, the US is not necessarily opposed to the general content of many of these
international treaties, and indeed that the US is interested in their existence as a source of
norm development.
Classic domestic explanations
The classic response to these puzzles is that the procedural rules in the advice and
consent process enables 1/3 of senators, or the senate foreign relations chair to block
treaties. This argument recognizes that domestic politics and institutions play an
important role in foreign policy (Allison 1971 ; Stoll 1984 ; Putnam 1988). Key to these
arguments about two-level games is that domestic veto players can credibly threaten to
withhold support for ratification (Mansfield, Milner et al. 2005). This is why Haller and
Holden have argued that a supermajority ratification rule such as the one in the senate
increases the bargaining position of a country, but it does so at the cost of also increasing
the risk that a treaty fails to become ratified (Haller and Holden 1997). In an analysis of
5
President's Committee on the International Labor Organization, Esther Brimmer, Assistant
Secretary, Bureau of International Organization Affairs, Department of Labor, Washington, DC, May 4,
2010 http://www.state.gov/p/io/rm/2010/141482.htm
5 ILO convention ratification by 17 countries, Boockmann also found that partisan
preferences of government parties influence ratification probabilities significantly, but
only when one considers the power allocation within the governing body (Boockmann
2006). Thus, a major obstacle to the Senate treaty process is the political opposition to a
treaty in the senate (Krutz and Peake 2009).
However, even traditional understandings of partisanship and domestic politics leave
Senate advice and consent behavior as a puzzle. For example, why do some treaties have
67 votes in support on the Senate floor as well as the support of the president and Senate
foreign relations chair, yet do not get brought up for a vote? The Law of the Sea clearly
meets this threshold. Others appear to have equally strong support. For example, the
Rotterdam Convention and the Stockholm Convention each made it out of committee, but
although both Republican and Democratic foreign relations committee chairs have
supported them, recommending their ratification, they have failed to move forward. Thus,
many treaties appear to have broad domestic political support, signaling that politicians
generally assess them as being in the interests of the US, yet they linger.
There is also a range of possible legal reasons that the US does not ratify treaties that it
generally supports. One such reason is that the US may simply not derive any added
benefit from ratifying it. This point has been made specifically regarding human rights
treaties, where the stability of the US constitution and the existence of a strong and
independent judiciary already guarantees citizens protection in many areas and where it is
not necessary to actually ratify the treaty for a state to obtain its benefits (Bradley 2010
331, 333). Relatedly, some treaties may stall because they conflict with the federal
structure of US policy making and related concerns about whether policy making power
resides at the federal or state level (Bradley 2010 330). Still, these explanations tend to be
very idiosyncratic to each treaty situation.
As with the basic national interest arguments, clearly the legal and basic institutional
arguments may often go a long way towards explaining the US behavior towards
international treaties. However, we propose another explanation that may be at work in
addition to these factors and may help explain the puzzling outcomes.
Opportunity costs
This paper presents an opportunity cost theory of treaty ratification. We argue that the US
sometimes delays or derails treaties not because they engender such opposition that
passing them would be impossible given the institutional rules, but because the President
and the Senate often does not prioritize them because they consider the opportunity cost
of doing so too costly. This occurs because the legislative agenda is a fixed and highly
prized resource.
Advice and consent as a two-step process
Before elaborating the opportunity cost theory, it is important to consider the treaty
ratification process itself. Indeed, the fate of US treaties, and thus an opportunity cost
6 theory of treaty ratification, cannot be examined without considering the process more
fully. After the US signs an agreement, the president decides whether to transmit it to the
Senate, and then the Senate Foreign Relations Committee (SFRC) decides whether to
recommend the treaty and pass it out of committee along with a treaty document drafted
by that Committee. This treaty document then moves to the floor for debate, possible
amendments, and a vote. To provide its advice and consent, the Senate must pass the
document with no less than a two-thirds majority. Only then can the president deposit the
instrument of ratification, formally ratifying the treaty.6
The nature of the two-step process means that it is necessary to account for behavior at
both stages to understand outcomes. We therefore theorize (and model) the treaty advice
and consent process as a two-step process. The first step, the transmittal decision, is
important to consider, because it may well be strategic. Indeed, when President Clinton
wanted to introduce the Convention on the Rights of the Child in 1995, Senator Jessie
Helms, Chair of the SFRC, felt so strongly about it that he, along with 26 cosponsors,
introduced a Senate Resolution urging the President to not transmit the Convention to the
Senate. Thus, if some Senators are very opposed to a treaty, they have incentives to voice
this opposition, which can be distracting or politically harmful for the president. In
addition, because the president usually endorses the treaty in the transmittal letter, the act
of transmittal is a request to the Senate for support for the endorsement. Thus, there may
be a reputational cost to transmitting treaties and have them stall (Krutz and Peake 2009
140). For several reasons, withholding transmittal can therefore be politically astute. Thus,
when theorizing the senate advice and consent process, it is important to consider both
the factors that influence the president’s transmittal decision and the senate’s advice and
consent process.
The fixed political agenda space and policy priorities
Opportunity costs arise when resources are fixed. We argue that political agenda space is
such a fixed resource; there are only so many things a president can do, and only so much
Senate floor time. When prioritizing legislative agenda time, Senators and presidents
favor issues that increase their popularity and chances for reelection (Mayhew 1974).
This usually means that senators focus foremost on domestic politics, which usually
drives elections (Abramson, Aldrich et al. 1987). This is not because the public does not
care about international issues. Indeed, the US public tends to care deeply about foreign
policy issues that have direct effects such as war and trade, and as a result, Senators also
devote considerable time to these. Indeed, trade agreements can engender considerable
domestic debate, as in the case of NAFTA, and when the US engages in war, the public
clearly cares deeply (Howell and Pevehouse 2007). However, these issues are not
typically the topics of treaties submitted for Senate advice and consent. Trade related
treaties are usually handled through congressional-executive agreements, and although
some article II treaties may be related indirectly to war, issues of specific wars are usually
not treaty issues.
6 In addition, at any time a treaty can be returned to the president or, if the committee
reports the treaty but no vote transpires, the treaty can go back to the committee. 7 Thus, Article II treaties are rarely the bread and butter of domestic politics. Voters may
hold preferences on a given treaty, but treaties rarely motivate voters. For example, the
US public has largely supported the International Criminal Court created by the Rome
Statute. However, confessed support for the ICC in treaty has not translated into a
mobilized voter base for the Rome Statute, and the issue is therefore not one that Senators
consider integral to their re-election.
The implication of this insight is central: the decision to push a treaty through the advice
and consent process may be less about an objective examination of costs and benefits of
the treaty (as assumed by traditional analysis) than about the benefit of spending time on
the treaty relative to the benefit of other possible agenda activity that may produce
important domestic legislation such as health care reform, for example. In other words:
senate advice and consent depends on the associated opportunity costs. The opportunity
cost can manifest itself for different types of treaties. Even non-divisive treaties require
some presidential attention to move through the process (Johnson 2010), and therefore
these may simply fall by the wayside because they gather little attention. Other, more
important treaties might also be affected by the opportunity cost, however. Even if
opponents might not technically command the requisite 1/3 of senators to block the treaty,
they may impose such high time costs that make the Senate and president are reluctant to
send sending onto the agenda of the Senate when there are many competing priorities.
So how great is the opportunity cost of treaty ratification in terms of presidential and
senate time? First and foremost, the opportunity cost is a function of the value of the
legislative agenda space at any given point: It depends on what else could be
accomplished with the time.
This theory leads to a quite unconventional proposition. Whereas pure institutional
analysis would expect the president to be most likely to transmit treaties to the senate
when he has a lot of political support and can garner the requisite 2/3 vote, the
opportunity cost theory predicts that this is exactly the time when he will be least likely to
spend time on treaties. When the president commands support in the Senate, the chances
of passing valuable legislation is higher (McCubbins and Cox 1993). In other words,
when the president commands support in the Senate, there is a surplus of political capital
and thus opportunity to move valuable legislation on the political agenda and the calendar
of the Senate. This boon of political capital raises the opportunity cost of the Senate
agenda time. Thus, the greater the president’s political support, the less likely he will be
to transmit treaties to the Senate.
In addition, although the House does not have a formal role in the treaty process, it does
have a role in the domestic legislation process. Because of this, we expect the level of
political support in both the House and the Senate should be relevant in the decision to
transmit. That is, with sizable majorities in both legislative chambers, the President
would be especially unlikely to transmit treaties to the Senate.
This leads to two observable propositions:
8 P1: The more co-partisans the president has in the House and the Senate, the less likely
he will be to transmit treaties to the senate.
P2: The greater the president’s approval ratings, the less likely he will be to transmit
treaties to the senate.
In the Senate, the value of floor time is more based on the political cycle. When the
president transmits a treaty the time costs rarely materialize in the very near term,
because the treaty has to pass through the SFRC process, which takes some time.
However, if the Senate places a treaty on the legislative calendar, then the impending
time cost is fairly immediate. This means that the Senate can be more sensitive in the
actually scheduling of treaties. Our opportunity cost theory predicts that the Senate will
be less inclined to spend time on treaties during times when Senators want to instead
push for domestic legislative victories.
Time is always scarce in the legislative branch (Hall 1996), but there is variation in that
scarcity. In election years, but particularly in presidential election years, domestic
legislation becomes too contentious to pass, particularly as the election grows nearer.
Indeed, it is expected that as the election season begins, the president and Congress do
not want to take major policy positions that could upset the electorate in the home
district.7 This lowers the cost of Senate time. This is especially true when the president is
not up for reelection, because the Senate is considered rather powerless to enact domestic
legislation at that time. Thus, calendar time becomes more plentiful. This leads to the
following expectation:
P3: In presidential election years, and especially during “lame duck” years, treaty
ratification will be faster.
P4: The senate is more likely to spend time on the treaty advice and consent process in
the fall of election years, than during the rest of the year.
Finally, the opportunity costs for the senate is greater when it has to vote on not only the
treaty itself, but also consider implementing legislation. The president often states in the
transmittal letter whether the treaty requires implementing legislation, and such
requirements will demand that the senate spends additional time on the implementing
legislation. In a busy senate, our opportunity cost theory would therefore expect
requirements for implementing legislation to impose additional delays or cause the treaty
to be set aside.
P5: Treaties that require implementing legislation will take longer to receive Senate
advice and consent.
7
Recall much of the controversial negotiations about the debt ceiling revolved around whether there would
be another debate that would be needed closer to the 2012 elections, with the clear implication that no
rational Congressperson would want to engage in that debate in the shadow of an election.
9 Revisiting other factors
As noted in the discussion already, standard political institutional analysis expects that
the Senate will not give advice and consent to treaties when 2/3 of Senators or the SFRC
chair oppose the treaty. As a matter of general expectation, Republican politicians should
be expected to be more opposed to treaty commitments. Republican Senators traditionally
stress sovereignty and conservative morals, both of which are elements that are at times
seen as threatened by multilateral agreements. For example, past work suggests that
Republicans are more reluctant to support arms treaties (Krepon and Caldwell 1991 ;
Delaet and Scott 2006). Scholars have also found that “conservative Senates and pivotal
senators systematically exploit the right to add reservations” (Auerswald and Maltzman
2003 1097), suggesting that conservative Senators are more cautious of offering their
advice and consent.
In general the greater tendency for more conservative politicians to oppose international
treaty commitments thus leads us to expect in general that democratic presidents will be
more inclined to transmit treaties, and that the more conservative the SFRC chair and the
2/3 senator the slower treaties will progress through the process.
However, actual support for passage may not be the only concern. Decision-making rules,
veto points, and legislative procedures open up the opportunity for small groups of
Senators to continue debate nearly indefinitely (Moravcsik 2005 150-151). This means
that a small group of the most conservative senators can essentially block a vote although
they do not necessarily have sufficient votes to defeat it were it brought to a vote. Based
on the opportunity cost theory, such opposition is important because it magnifies the
expected time the president or senator foreign relations chair would expect debate on the
treaty to take and thus may deter the president or senate from action. Senator Richard
Lugar (R-Ind) lamented this state of affairs in a 2009 speech to the Washington Foreign
Law Society:
“[T]oo often, narrow objections to treaties are allowed to prevent any
Senate consideration. The Senate’s rules allow the Senate to vote to cut off
debate on a treaty, and thereby to ensure that a vote on the treaty may
occur. But the Senate’s leadership – both Democrat and Republican -- has
been reluctant to exercise this option in recent years. Their concern
appears to be that doing so would require the Senate to spend too much
time debating particular treaties, to the exclusion of other matters that have
greater domestic political appeal. As a result, the Senate has not held a
significant debate on a treaty since the late 1990s, when the Senate
debated and approved the Chemical Weapons Convention, and debated
and rejected the Comprehensive Test Ban Treaty. In essence, for more
than a decade, treaties have had to meet a standard for Senate approval
10 that approaches unanimity -- far exceeding the already difficult two-thirds
majority requirement established in the Constitution.”8
In sum, based on both the opportunity cost theory and on a simple preference-based
theory, the general expectation is that the more opposed Senate veto players, including
the most conservative senators, are to a treaty, the longer the Senate advice and consent
process will take.
As also noted, there are many legal properties, which may explain the fates of particular
treaties, just as treaties on certain issues may engender unique lobbies for or against the
treaties. Because our goal here is not to explain what all these treaty properties may be,
this analysis will not include highly treaty specific attributes. However, it does consider
the general subject matter of treaties and expects, in particular, that human rights treaties
will take longer to progress through the advice and consent process.
Descriptive data
Time period
For the statistical analysis, our data includes multilateral treaties in the advice and
consent process post-1967. This means that the treaties may have been signed prior to
1967, but they are only in our sample if they had not yet been ratified by then. Thus, they
have to be still untransmitted or pending, by 1967.9 Notably, we include some treaties
that the US has signed, but the president has not yet transmitted them to the Senate.
Assembly a list of these treaties is difficult, because no list of them exist. Furthermore, it
is potentially tricky to decide which international agreements belong on this list, as some
of them may be executive agreements that do not require senate advice and consent. To
devise an accurate list, we searched multiple available treaty databases for agreements
signed by the United States, but to which the United States had not yet become a party.
These databases include the United Nations Treaty Database, ECOLEX,10 The Hague
Conference on Private International Law,11 and Oceana Law.12 We also compared our list
against a list of multilateral treaties from Gamble’s dataset on treaties and Wikipedia’s
“treaties by year of conclusion.”13 After assembling this list, we consulted with the Office
of the Assistant Legal Adviser for Treaty Affairs in the State Department. Here Frank
Holleran, the Treaty Analyst and Depositary Officer, helped us identify treaties that were
not designated to go through the advice and consent process, and also added some treaties
that we had missed. Thus, we are quite confident that our list of un-transmitted treaties is
8
(Lugar 2009)
go back to 1945 to look for unperfected treaties which could still be on the table as of 1967. http://www.ecolex.org/start.php
11
http://www.hcch.net/index_en.php
12
http://www.oceanalaw.com/
13
http://en.wikipedia.org/wiki/Category: Treaties_by_year_of_conclusion
9 We
10
11 fairly accurate and that the treaties included are indeed ones requiring senate advice and
consent.
Bilateral treaties
We limit our statistical analysis to multilateral treaties, for several reasons. About half the
treaties that go through the advice and consent process in the United States are bilateral,
but these bilateral treaties differ from multilateral treaties in at least four important ways.
First, bilateral treaties tend to deal with a narrow set of topics. For example, the bilateral
treaties transmitted to the senate between 1967 and 2011 belong primarily to five
categories: consular conventions (17), property or copyright treaties (9), tax conventions
(129), investment treaties (49) and extradition treaties (147). Together these treaties make
up 350 treaties, which constitute 88 percent of all the non-multilateral treaties in Thomas
that were transmitted in 1967 or later.
Second, bilateral treaties often are renewals or revisions of treaties with a given country
repeatedly. Thus, for example, 8 of the tax conventions approved for senate advice and
consent since 1967 are with Canada. This means that these treaties often are minor
revisions of previous documents, which reduces the need for debate. For these reasons,
Galbraith calls these treaties “repetitive bilateral treaties” (Galbraith 2012 4).
Third, bilateral treaties often follow various standard formats. This enables the senate to
deal with them in bundles, meaning that the senate passes two or more treaties of the
same kind on the same date, and often as part of the same vote. Typically 4-8 treaties are
bundled. Thus, half the consular conventions were bundled with at least one other treaty.
The senate bundled 95 percent of the tax conventions, 89 percent of the property or
copyright treaties, 87 percent of extradition treaties and 88 percent of investment treaties.
The senate bundled extradition treaties 82 percent of the time, in one case providing
advice and consent to 30 treaties on the same day. This bundling means that any delay in
passing these bilateral treaties is irregular because the senate will hold them until there is
a hearing for treaties of a given type. Indeed, the senate routinely holds hearings on
treaties of a given type once a year or every two years (interview). Thus the standard
format of most bilateral treaties leads the senate to deal with them very differently than
multilateral treaties which must by and large by dealt with on an ad hoc basis.
Thus, bilateral treaties differ significantly from multilateral treaties, both in content,
partners, and in how the senate chooses to handle them. That said, the fact that the senate
indeed chooses to bundle these treaties provides support for the opportunity cost
argument. The similarity of these treaties reduces the need for debate and subsequently
facilitates bundling, which in turn reduces the need for senate floor time and thus the
opportunity cost to the senate.
Congressional Executive Agreements
It is for similar reasons that we do not include executive agreements in the analysis,
although for purposes of international law they are identical. Executive agreements take
12 two forms: sole and congressional. Sole executive agreements are made by the president
alone and actually constitute the vast majority of international agreements to which the
US is a party. Sole-executive agreement can only cover matters within the President's
authority or matters in which Congress has delegated authority to the President. With a
few notable exceptions, these agreements are routine. Congressional executive
agreements are another matter. These require passage in both the House and the Senate.
However, the requirement for passage is only a majority vote. Therefore the possibilities
for blocking the legislation and taking up valuable floor time are not as great, and the
opportunity cost of pushing the agreement to a vote is therefore not as great either.
Furthermore, the congressional executive agreements tend to concern matters of trade,
the one foreign policy area outside of war that probably concerns citizens most because it
influences both jobs and consumer prices (Fordham and McKeown 2003). Thus, we
would expect legislators to be more willing to spend time on congressional executive
agreements.
Multilateral treaties
As discussed above, we focus on multilateral treaties for several reasons. In addition to
the concerns above, a focus on multilateral treaties also has the benefit of focusing the
analysis on those agreements most likely to influence broader global cooperation.
If we limit the sample to multilateral treaties in the advice and consent process post-1945,
there are 394 treaties for which we have complete information. Of these, the president
eventually transmitted 354, and the average time to transmittal is about 3.2 years
(standard deviation about 6.5 years), with the maximum being 49 years. Of the treaties
transmitted, 296 were ultimately ratified in an average of 1.9 years, with the standard
deviation being 4.4 years and the longest spell being 41 years
For data visualization and analysis purposes, however, we limit our sample to post-1967
treaties. The next three figures display absolute signature, transmittal and ratification
trends over time, but only for treaties signed after 1967. Figure 1 displays signatures.
Note the bursts of signatures in the early 1970s and the 1990s, and also the much lower
rate of US signature of international agreements in recent years: Whereas until 2003 the
US has always signed at least 4 multilateral treaties a year, after 2003 there has been
three years when it signed only one such agreement, and one year, 2008, when it signed
none. This shows that the flow of treaties available for transmittal is not at all constant,
something we will consider later in the model. Figure 2 shows that the number of
transmitted treaties likewise varies a lot from year to year. It is interesting to both note in
Figure 2 and 3 the low activity during the Obama administration. Indeed, the year 2010
under president Obama marks the first year in our sample that no multilateral treaty was
ratified at all.
13 15 10 5 0 Figure 1: Number of multilateral treaties signed by the United States per year since 1967
18 16 14 12 10 8 6 4 2 0 Figure 2: Number of multilateral treaties transmitted to senate per year since 1967
16 14 12 10 8 6 4 2 0 Figure 3: Number of multilateral treaties receiving advice and consent by the senate per
year since 1967
Our hypotheses expected that treaty ratification would be faster in presidential election
years, and especially during lame-duck years. This means that we should expect to see
14 more multilateral treaties pass during those times. Figure 4 shows how many multilateral
treaties pass on average in a particular type of year. It is clear that on average the fewest
multilateral treaties are passed in non-election years, more in midterm election years, and
even more in presidential election years. Finally, the average number of treaties passed in
“lame duck” years is clearly the highest, and is more than twice that in non-election years.
12 10 8 6 4 2 0 non election years Midterm election year presidential lameduck election election years years Figure 4: Average number of multilateral treaties given senate advice and consent during
a particular type of year, 1967-2011
It is also interesting to consider the activity in the Senate throughout the year, given that
we have expectations that the fall, especially before elections, will tend to be more active
for treaties. Figure 5 shows the frequencies of treaty ratification in months throughout the
different types of years. In general there is a certain rhythm to the year. Even in nonelection years, treaties are more likely to be passed late summer and fall. However this
trend gets increasingly pronounced as we move into election years, then presidential
election years and finally “lame-duck” years, where it is by far the most pronounced with
September seeing a virtual flurry of treaties.
15 lameduck years pres election years (not lameduck) midterm election years non election years December November October September August July June May April March February January December November October September August July June May April March February January December November October September August July June May April March February January December November October September August July June May April March February January 0 1 2 3 4 5 6 7 Figure 5: Average number of multilateral treaties passed in a given month, 1967-2011
16 Analysis
Note that we have cast our theory as one of timing: although Presidents do refuse to send
treaties to the Senate and the Senate does occasionally vote down a treaty (the last time as
the senate rejected a treaty was the Comprehensive Nuclear Test Ban Treaty in 1999), the
more interesting question is to understand what factors lead to delay in the Senate advice
and consent process.14
Moreover, because at least some presidential transmittal decisions are likely to be
strategic, simply examining all treaties that are transmitted leaves out a valuable and quite
likely non-random set of the sample. On the other hand, simply modeling signature time
to the advice and consent decision oversimplifies the process and ignores that there are
some treaties the Senate cannot provide with its advice and consent, because the president
has never transmitted them to that body.
We therefore first model the transmittal decision, so that our ratification model can take
the transmittal decisions into account. As noted, treaties may fail at the stage of
presidential transmittal. More commonly, as previously discussed, a president may hold
a treaty for some time before transmitting – we chose to explicitly model this decision as
well as the choice of the Senate to provide advice and consent.
To analyze our hypotheses as a part of this two-stage process, we undertake a two-step
multivariate analysis of our treaty data. We first estimate a model for the transmission
decision, then follow with a model based on the advice and consent decision. Our data is
time-series, cross-sectional by nature, set up in treaty-year format, to capture changes in
the key independent variables over time, should a treaty remain on the president’s desk or
in the senate for more than one year. Thus, each variable we describe below is coded on
a yearly basis.
Core Variables
Given our arguments above, we expect several variables to influence the length of time it
takes for Presidents to transmit treaties to the Senate as well as the time from transmittal
until the senate provides advice and consent.
We examine the influence of political party support in both houses of the legislature (P1)
by defining President % Control, which is the percentage of the House and Senate
controlled by the party of the President. Traditionally, scholars would expect this variable
to lead to an easier road to advice and consent. We, however, expect this to be associated
with a longer time to advice and consent. We include this variable in both states (transmit
and advice and consent) of our model.
We expect higher presidential approval ratings (P2) to be correlated with a lower desire
to transmit treaties to the Senate, given that this approval is likely to translate into bigger
14
In a similar vein, scholars have chosen to analyze delays in Senate consent on nominations. See
(McCarty and Razaghian 1999 ; Binder and Maltzman 2002 ; Derouen, Peake et al. 2005)
17 momentum on domestic legislative imperatives. To examine this idea, we introduce
Approval, the annual average of the Gallup presidential approval poll for each year of
observation.15 We include this variable only in the first stage of our model given that this
is a factor that presidents are likely to consider, but not the Senate in their decisionmaking.
To address P3 and P4, involving the dynamics of the electoral calendar and its effect on
opportunity costs, we define two variables: Lame duck and Re-election. The first of these
two variables is coded as 1 in the final year of a president’s second term. The second of
these is defined as 1 in the final year of a president’s first term, when he is up for reelection. We expect both of these variables to be expected with shorter advice and
consent times given that there is little incentive for the Senate to work on domestic issues
during these two periods of time. Moreover, we include these only in the ratification
portion of the model since they are likely to be more important for Senate decisionmaking.16
Finally, we address P5 by introducing the variable Implementation, an indicator variable
coded “1” if further domestic legislation is required for implementation of the treaty – i.e.
the treaty is not self-executing. The variable is coded “0” otherwise.
Control factors
Of course, as previously discussed other factors may influence both the president’s
decision to transmit a treaty to the Senate as well as the Senate ratification process. To
this end, we introduce several control variables to both models.
First, we define SFR Chair as the DW-Nominate score of the chair of the Senate Foreign
Relations Committee. The Nominate score measures the underlying liberal-conservative
spectrum of the voting record of members of Congress (Lewis and Poole 2004). Higher
Nominate scores suggest a more conservative committee chair, while lower values
suggest a more liberal chair. Because the chair of the committee remains for two years,
this variable varies every other year.
Second, we also use the DW-Nominate scores to define Conservative Senator to measure
the ideology of the most conservative Senator in the chamber. This easily observable
metric should influence both the President’s decision to transmit a treaty as well as the
length of time a treaty takes to move through the advice and consent process. We expect
that the more conservative the most conservative Senator, the longer both transmittal and
ratification will take.
Third, we define Democratic President, which equals 1 if the President is a Democrat.
Our expectation is that Democratic Presidents will be more likely to transmit sooner and
push for quicker advice and consent in the Senate.
15
Data is taken from the Gallup web site. www.gallup.com.
16 Inclusion of these variables in the transmittal stage of the model yields no evidence they influence presidential decision-­‐making. 18 Fourth, we calculate the number of treaties available to be transmitted based on the
number of untransmitted treaties left from previous presidents and the new number of
treaties concluded during the year. We label this variable Treaties Available for
Transmittal. We control for these same dynamics in the ratification stage by computing
the number of treaties under consideration in the Senate, labeled Treaties Available for
Advice and Consent.
Finally, it is also important to control for the type of treaty under consideration. Issue
area could have important implications for the speed at which the treaty moves through
the process. To this end, we classify treaties into four broad categories of types: Human
Rights (e.g., UN human rights treaties), Commercial (e.g., trade agreements),
International Law (e.g., technical treaties regarding rules and procedures in international
law), Environment (e.g., regulatory treaties involving the environment), and Arms
Control (e.g., multilateral arms treaties such as the Comprehensive Test Ban Treaty).
In terms of the ratification model, we also are interested in the effects of the first stage
(transmittal) on the second (advice and consent). This is investigated by calculating a
predicted time to transmit from the President’s desk. We estimate a first stage model,
then calculate a predicted median survival time – that is, the time the treaty will sit on the
president’s desk. Again, we expect that longer periods of time on the president’s desk
should lead to faster action in the second stage of the process: Senate advise and consent.
Our initial models thus hold that transmittal and advice and consent are a function of the
following:
Transmittal = f(President % Control; Approval; Type: International Law; Type: Human
Rights; Type: Arms Control; Type: Commercial; Democratic President; Transmit
Available; SFRC Chair; Conservative Senator)
Advice and Consent = f(President % Control; Type: International Law; Type: Human
Rights; Type: Arms Control; Type: Commercial; Democratic President; Advice and
Consent Available; SFRC Chair; Conservative Senator; Lame duck; Re-election;
Predicted Transmit Time)
To analyze data on time to transmittal or advice and consent, we turn to event history or
duration models following Krutz and Peake (2009). As shown in Figures 6 and 7, the
underlying distributions of the time-to-transmit and time-to-advice and consent
dependent variables follows a Weibull distribution. In our subsequent multivariate
analysis, we estimate a series of Weibull models, using the accelerated failure time metric
to ease interpretation. For our coefficient estimates, it should be noted that positive signs
indicate a positive effect on the hazard rate, meaning shorter transmission/ratification
times. Negative signs indicate longer transmission/ratification times.
Figure 6. Hazard rate for treaties signed by the US transmitted to the Senate.
19 0.00
0.25
0.50
0.75
1.00
Kaplan-Meier survival estimate
0
10
20
30
40
50
analysis time
Figure 7. Hazard rate for ratification of treaties transmitted to Senate.
0.00
0.25
0.50
0.75
1.00
Kaplan-Meier survival estimate
0
20
40
60
analysis time
20 Results
The estimates of the transmit stage of the model are found in column 1 of Table 1.
Consistent with our expectations, President Percent Control is negative and statistically
significant. This finding is strong evidence of the opportunity cost logic: the more copartisans the President has in Congress, the longer he takes to transmit treaties.17 Again,
this logic is in contrast to extant theories of the production of domestic laws, where it is
presumed stronger partisan support will facilitate legislative production. Moreover, this
effect is substantively important as well. Increasing the average number of co-partisans
by nine seats yields a increase in the predicted transmit time of over 35 percent.18
Similarly, the estimate of Approval is negative but not statistically significant, indicating
that higher levels of public approval for the president have little influence on transmit
times for treaties.
Several control variables do achieve statistical significance and the estimates are
consistent with our theoretical expectations. First, SFR Chair is negative and statistically
significant, indicating that more conservative chairs lead to a lower hazard rate – i.e.,
longer transmit times. Similarly, the more conservative is the most conservative Senator,
the longer the transmit time from the President’s desk, as suggested by the negative and
statistically significant coefficient on Conservative Senator. Conversely, the positive
estimate of Democratic President indicates that democratic presidents more quickly
transmit signed treaties to the Senate for ratification. The substantive effect of this
variable is quite high: switching from a Democratic to a Republican president yields a
predicted increase of transmittal time of nearly 2 years: a doubling over the baseline
prediction.
And while the Commercial, International Law, and Human Rights indicators for treaty
type are statistically distinct from the reference category, chi-square tests show they
cannot be distinguished from other treaty categories. The rest of our treaty type categories
are statistically indistinguishable from the reference category and each other. It is also
worth noting that if we remove these type variables completely from the model, our
estimates do not change significantly, suggesting that although these treaty specific
variables may be important in explaining the outcomes, they are not systematically
related to our variables of interest.
This first stage model is important in that it both helps to determine what makes the
president decide to send treaties into the advice and consent process, and it allows us to
generate an instrument for use in the second model, which estimates the time-toratification once the treaty hits the Senate.19 We thus use this transmit model to predict
17 We get a similar result if only the Senate is computed in this measure. 18 All predicted probabilities are computed against a baseline model assuming a democratic president, an environmental treaty, and all other variables held at their mean value. 19 Note that the model gains identification from the inclusion of Approval in the first stage and Lame duck and Re-­‐election in the second stage. In addition, because Treaties Available is 21 the time that a treaty remains with the president – and later introduce this predicted
variable into the second stage of our model.
We now have an instrument that contains the factors that will influence whether the
president believes a treaty will have difficulty in the Senate. Should he believe this is the
case, we would see longer expected transmit times. If the president’s strategic decisions
are correct, our prediction is that longer transmit times should predict shorter ratification
times, since the president should have waited for an opportune time to act.
Turning to our model of Senate ratification, the estimates of the ratification model can be
found in Column 2 of Table 1. Note that nearly all of the domestic political variables that
were statistically significant in the first stage do not achieve statistical significance in the
second stage. Namely, Democratic President, President Percent Control, and
Conservative Senator do not achieve conventional levels of statistical significance.
Rather, these domestic political factors exert their influence in the decision of the
President to transmit treaties.
SFR Chair, however, is statistically significant in this second stage. Thus, a more
conservative chair of the Senate foreign relations committee slows down the advice and
consent process, independently of its effect on delaying presidential transmission.
Increasing the conservativeness of the chair by one standard deviation leads to a
predicted increase of over 50 percent in the time of advice and consent.20
Lame duck is positive and significant, indicating that the Senate moves faster on treaties
in the final year of an executive’s eight year time in office when Senate floor time is less
valuable. This variable has a large substantive effect: moving from a lame duck year to a
non-lame duck year yields an over 70 percent increase in the time to advice and consent.
Similarly, Re-election is positive and significant, indicating that treaties move through the
advice and consent process about 40 percent faster in presidential re-election years.
Next, we re-estimate the model, but include the variable Predicted Transmit, which is the
predicted transmit time from the previously estimated model. The estimate of Predicted
Transmit is statistically insignificant, as shown in column 3 of Table 1. This indicates that
a longer predicted time of transmittal has no influence on the ratification process. One
interpretation of this result could be that presidents do not accurately read the strategic
ratification game in front of them. If holding a treaty longer led to quicker ratification
times, this would indicate successful strategic delay to speed the acceptance of the treaty.
We find no empirical evidence of this, however.
We extend our ratification model in two ways. First, we check an alternative version of
our ideology variable by measuring the NOMINATE score of the 67th Senator and using
this measure to replace Conservative Senator. The estimate of this new variable,
measured as only those treaties on the President’s desk in the first stage, it can also be used to help in model identification. 20 In the second stage models, the baseline probability is computed on an environmental treaty, democratic president, Re-­‐election set to 0, and Lame duck set to 1. 22 Senate67 is not statistically significant – thus, whether one examines the most
conservative end of the ideological spectrum in the Senate or the threshold vote for
passage, ideology does not appear to matter in a systematic way.
Second, as indicated by P5, it is possible that the requirement of implementing legislation
could slow down the treaty ratification process as Senators balk at the additional
requirement to change the status quo. To check for this possibility, we code an indicator
variable Implementation, which is coded as a “1” if the treaty will require additional
legislation, “0” otherwise.21 We add this new variable to our original ratification model.
The estimates of this new model can be found in column 4 of Table 1. The new variable
is statistically significant, although its inclusion does not change any of the major
substantive variables of interest. Still, this does suggest that treaties requiring further
legislative effort are more likely to experience a delay in ratification.
In sum, when considering these models, it is a combination of the effect of the ideology
of the committee chair, along with the electoral calendar that appear to drive the advice
and consent process. Few other political factors or treaty characteristics seem to influence
the timing of treaty advice and consent.
Conclusion
Existing international relations theory expects states to join treaty regimes if they serve
the national interest. Yet it does not take much to see this as an oversimplification: there
are many cases where the US has not ratified a treaty, yet shows its support for the treaty
in other ways, indicating that the treaty serves its interests. To explain this, scholars of
American politics will quickly add that, of course, what really matters is whether actors
with veto powers in the political process believe that the treaty is in the national interest.
Because of the power they hold over the process, the political preferences of the senate
foreign relations chair or the 2/3-senator can derail the senate advice and consent process.
These actors may oppose treaties based on principled issues that have little to do with the
substance of the treaty cooperation. Or strong lobbies against a treaty may deter them
from supporting the treaty. Whatever the origin of their preferences, they have the ability
to essentially hold the treaty hostage.
We agree that is not possible to look at the national interest in ratifying a treaty in
isolation from the domestic political opposition. Indeed, we clearly show the importance
of veto players. However, we also argue that not even this traditional veto-player analysis
fully explains US treaty participation. Sometimes those opposed to the treaty are not in
veto positions. Sometimes even benign treaties, with little opposition at all, fall by the
wayside.
Our analysis reveals another much less appreciated factor that influences the treaty
process: the opportunity costs of Senate time. 21 The sample size of this model is smaller as we are missing data on a number of treaties concerning their requirement of implementing legislation. We are currently in the process of filling in this missing data. 23 Our analysis has shown that the United States often delays or abstains from ratifying
treaties because the constitutionally required advice and consent process increases the
political costs of treaty ratification. Specifically, we have articulated an opportunity cost
theory of treaty ratification. This theory argues that the fixed agenda space of the Senate
adds an opportunity cost to ratification, because the president and the senate has to decide
whether to spend time on the treaty or to spend time on other, often more valued,
domestic legislation. The opportunity cost is partly a function of political preferences, but,
more interestingly, it is also a function of the political value of the Senate floor time.
Thus, the decision to push a treaty through the advice and consent process may be less
about an objective examination of costs and benefits of the treaty (as assumed by
traditional analysis) than about the benefit of spending time on the treaty relative to the
benefit of other possible agenda activity that may produce important domestic legislation
such as health care reform. In other words: Senate advice and consent depends on the
associated opportunity costs.
This insight is what explains otherwise perplexing patterns of US treaty behavior such as
the fact that Senate support for the president delays rather than accelerates treaty
ratifications. However, this pattern becomes understandable in the light of our
opportunity cost theory: The more support the president enjoys, the more productive the
Senate is being in passing domestic legislation, and the further the president and the
Senate are from their re-election campaigns, the more valuable the time of the Senate
agenda is. Thus, these conditions increase the opportunity costs for treaties, and the
president or the senate may find them so large that the treaty either gets delayed or
becomes entirely stuck.
These findings speak both to American politics and international relations. For American
politics, the findings reveal something new and interesting. Prior analysis of the other
Senate advice and consent process, namely that of presidential nominations, has found
that divided government, political polarization and elections obstructs Senate advice and
consent. In contrast, we find that treaty advice and consent is more, not less, likely when
the president lacks strong support in the Senate, when the Senate is gridlocked, and
during election years. Based on our opportunity cost theory, we argue that this is because
presidential nominations are domestic matters, which are forced onto the senate agenda
whereas treaties are foreign policy matters, which are optional and face multiple hurdles.
Indeed, the work on presidential nominations demonstrates the core of the opportunity
cost argument, namely that it is time consuming to force controversial items on the
Senate agenda.
For international relations theory, the findings mean that there is not as straightforward
translation from some unified national preferences into national politics as traditional
cross-national analysis of treaty ratification assumes. Traditional analysis of state
participation in international treaties usually treats states as unitary actors and assumes
that their decision to join treaties depends purely on an analysis of its costs and benefits
and that states therefore will join treaties that are in their interests. More sophisticated
incorporation of domestic preferences as advocated by theory on two-level games is
24 useful, but it too expects that treaties that enjoy wide support would receive timely Senate
advice and consent. However, we have found that in the United States, senate rules can
make treaty ratification time consuming if there is even the slightest opposition to a treaty.
Importantly, a particular treaty may simply be too low a priority to compete with other
legislative demands. Thus, even treaties that enjoy the support of 2/3 of Senators, enough
technically to win senate advice and consent, may have to wait a long time to be
scheduled for a floor vote or may never make it.
This finding also has implications for global cooperation. The United States often plays a
pivotal role in whether international cooperative efforts are successful. Without the
United States, it may be harder to get other countries to join, and the United States itself
may lack a full voice within newly created organizations.
25 Table 1: Weibull models of the multilateral treaty process, 1967-2008.
Ratify
-1.625
(1.028)
--.--
Ratify
-1.455
(1.163)
--.--
Ratify#
-1.630
(1.063)
--.--
0.182
(0.234)
-1.497***
(0.376)
0.125
(0.320)
0.194
(0.242)
-0.056
(0.197)
0.004
(0.012)
-0.610***
(0.165)
0.092
(0.717)
--.--
0.199
(0.249)
-1.459***
(0.401)
0.114
(0.321)
0.242
(0.286)
-0.093
(0.234)
0.004
(0.012)
-0.582***
(0.190)
--.--
Predicted Transmit
Transmit
-3.535***
(0.904)
-0.007
(0.004)
-0.345*
(0.186)
-0.632**
(0.298)
0.001
(0.233)
-0.505***
(0.174)
0.698***
(0.169)
0.006
(0.007)
-0.627***
(0.148)
-1.392*
(0.740)
--.--
Lame duck
--.--
Re-election
--.--
Implementation
--.--
0.636***
(0.180)
0.465***
(0.121)
--.--
-0.024
(0.085)
0.642***
(0.181)
0.463***
(0.122)
--.--
0.302
(0.264)
-1.296***
(0.411)
0.438*
(0.237)
0.347
(0.238)
0.221
(0.190)
-0.014
(0.012)
-0.814***
(0.160)
0.691
(0.713)
--.--
Constant
1.259**
(0.630)
-0.960***
(0.810)
-1.049
(0.840)
0.374**
(0.168)
0.334***
(0.114)
-0.349*
(0.198)
-0.848
(0.844)
N
ln_p
1815
-0.028
(0.044)
1653
-0.254***
(0.058)
1653
-0.255***
(0.058)
1334
-0.107*
(0.057)
President % Control
Approval
International Law
Human Rights
Arms Control
Commercial
Dem. President
Availability+
SFR Chair
Conserv. Senator
Note: *** = p<0.01; ** = p<0.05; * = p<0.10; all tests two-tailed. Robust standard error
estimates are computed. + = Availability altered based on transmit or ratification stage.
#
This model limited to treaties signed after 1967.
26 REFERENCES Abramson, P., J. Aldrich, et al. (1987). "Progressive Ambition among United States Senators: 1972ñ1988." The Journal of Politics 49(1): 3-­‐35. Allison, G. (1971). Essence of decision: explaining the Cuban missile crisis. Boston, Massachussetts, Little Brown. Auerswald, D. and F. Maltzman (2003). "Policymaking through advice and consent: Treaty consideration by the United States Senate." Journal of Politics 65(4): 1097-­‐1110. Binder, S. and F. Maltzman (2002). "Senatorial delay in confirming federal judges, 1947-­‐1998." American Journal of Political Science 46(1): 190-­‐199. Boockmann, B. (2006). "Partisan politics and treaty ratification: The acceptance of International Labour Organisation conventions by industrialised democracies, 1960–1996." European Journal of Political Research 45(1): 153-­‐180. Bradley, C. (2010). "The United States and Human Rights Treaties: Race Relations, the Cold War, and Constitutionalism." Chinese Journal of International Law 9(2): 321-­‐344. Buthe, T. and H. Milner (2008). "The politics of foreign direct investment into developing countries: increasing FDI through international trade agreements?" American Journal of Political Science 52(4): 741-­‐762. Charnovitz, S. (2008). "Editorial Comment: The ILO Convention on Freedom of Association and Its Future in the United States." American Journal of International Law 102(1): 90-­‐107. Delaet, J. and J. Scott (2006). "Treaty-­‐Making and Partisan Politics: Arms Control and the U.S. Senate, 1960–2001." Foreign Policy Analysis 2(2): 177-­‐200. Derouen, K., J. Peake, et al. (2005). "Presidential mandates and the dynamics of Senate advice and consent, 1885-­‐1996." American Politics Research 33(1): 106-­‐131. Elsig, M., K. Milewicz, et al. (2012). Who is in love with multilateralism? Treaty commitment in the post-­‐Cold War era. Evans, L. and W. Oleszek (2003). A Tale of Two Treaties: The Practical Politics of Treaty Ratification in the US Senate. Congress and the Politics of Foreign Policy. C. Campbell, N. Rae and J. Stack Jr. Upper Saddle River, New Jersey, Prentice Hall: 70– 89. Fordham, B. and T. McKeown (2003). "Selection and Influence: Interest Groups and Congressional Voting on Trade Policy." International Organization 57(03): 519-­‐549. Fredriksson, P. (2000). "Ratification of the 1992 climate change convention: what determines legislative delay." Public choice 104(3/4): 345. Galbraith, J. (2012). "Prospective Advice and Consent." Yale Journal of International Law Forthcoming. Hafner-­‐Burton, E. (2009). Forced to be Good: Why Trade Agreements Boost Human Rights. Ithaca, Cornell University Press. Hall, R. (1996). Participation in Congress. New Haven, Yale University Press. 27 Haller, H. and S. Holden (1997). "Ratification requirement and bargaining power." International Economic Review 38(4): 825-­‐851. Johnson, T. (2010). Interview with John B. Bellinger III, U.S. Trouble with START and Other Treaties, July 9. Krepon, M. and D. Caldwell (1991). The Politics of arms control treaty ratification. Hampshire, United Kingdom, Palgrave Macmillan. Krutz, G. and J. Peake (2009). Treaty politics and the rise of executive agreements: international commitments in a system of shared powers. Ann Arbor, University of Michigan Press. Lewis, J. and K. Poole (2004). "Measuring Bias and Uncertainty in Ideal Point Estimates via the Parametric Bootstrap." Political Analysis 12(2): 105-­‐127. Lugar, R. (2009). "Speech on treaties to the Washington Foreign Law Society, 14 September." Retrieved October 14, 2010, from http://lugar.senate.gov/news/record.cfm?id=317861&&. Mansfield, E., H. Milner, et al. (2005). "Vetoing cooperation: The impact of veto players on international trade agreements." American Political Science Association, Washington, DC. Martin, L. (2005). "The President and International Commitments: Treaties as Signaling Devices." Presidential Studies Quarterly 35(3): 440-­‐465. Mayhew, D. (1974). Congress: The electoral connection. New Haven, Connecticut, Yale University Press. McCarty, N. and R. Razaghian (1999). "Advice and consent: Senate responses to Executive branch nominations 1885-­‐1996." American Journal of Political Science 43(4): 1122-­‐1143. McCubbins, M. and G. Cox (1993). Legislative Leviathan. Berkeley, California, University of California Press. Moravcsik, A. (2000). "The origins of human rights regimes: Democratic delegation in postwar Europe." International Organization 54(2): 217-­‐252. Moravcsik, A. (2005). The Paradox of US Human Rights Policy. American Exceptionalism and Human Rights. M. Ignatieff. Princeton, New Jersey, Princeton University Press: 147-­‐197. Putnam, R. (1988). "Diplomacy and domestic politics: the logic of two-­‐level games." International Organization 42(3): 427-­‐460. Simmons, B. (2000). "International Law and State Behavior: Commitment and Compliance in International Monetary Affairs." American Political Science Review 94(4): 819-­‐835. Simmons, B. (2009). Mobilizing for Human Rights: International Law in Domestic Politics. New York, New York, Cambridge University Press. Stoll, R. (1984). "The guns of November: Presidential reelections and the use of force." Journal of Conflict Resolution 28(2): 231-­‐246. Vreeland, J. (2008). "Political institutions and human rights: Why dictatorships enter into the United Nations Convention Against Torture." International Organization 62(1): 65. 28